# Mikita Shylai > Mikita Shylai is a product engineer and founder who builds mobile and web software and applies AI to practical workflows, with a strong interest in the business and economics behind products. Source: https://shyl.ai. Last substantive update: 12 October 2026. Only published, publicly approved information is included. Metrics carry the date they were measured. ## Summary - Name: Mikita Shylai - Role: Product engineer and co-founder - Website: https://shyl.ai - LinkedIn: https://www.linkedin.com/in/shylai/ - GitHub: https://github.com/MKraKoff I build products people use, and I study the economics that decide whether they last. I usually start in the code, but the questions I find most interesting are about distribution, pricing and incentives: the things that decide whether a useful product lasts. ## Key facts - 50K+ users and $50K+ revenue (OpenRoad · co-founder · 2026). Record: https://shyl.ai/projects/openroad - Youngest Google for Startups resident (Campus Warsaw · Codaro · 2025). Record: https://shyl.ai/awards#google-for-startups-residency - 1st place at TeenCrunch (Poland’s largest teen accelerator · 2025). Record: https://shyl.ai/awards#teencrunch - 9th in Poland, economics olympiad (from zero in nine months · 2026). Record: https://shyl.ai/awards#owe-xxxix ## Background Mikita Shylai builds software products end to end, from mobile apps to AI-assisted document and knowledge tools, working in TypeScript, React and Next.js, Python and C#. Alongside engineering, Mikita studies economics and entrepreneurship and treats them as part of the same job: understanding who a product is for, how it reaches them, and why it is worth paying for. ## Areas of work - **Engineering:** Shipping mobile and web products end to end, with care for the parts users never see: billing, data and reliability. - **Applied AI:** Retrieval and document workflows where every answer can be traced back to its source. - **Business:** Distribution, pricing and the small decisions that turn a working product into a sustainable one. - **Economics:** Incentives, trade-offs and market structure, as a way to reason about products and the markets around them. ## How I work ### Mobile, AI agents, cloud and blockchain, each in a real product. Native iOS and Android. Five AI agents that publish legal articles with no human in the loop. Document Q&A that cites every answer. Smart contracts. Google Cloud backends. Evidence: [OpenRoad](https://shyl.ai/projects/openroad), [TaxCourtBlog](https://shyl.ai/projects/taxcourtblog), [Cited](https://shyl.ai/projects/cited), [PrivateBond](https://shyl.ai/projects/privatebond), [Codaro](https://shyl.ai/projects/codaro) ### I follow traction, not my first idea. The best founders spend their time with customers, not on their first idea. Codaro and mTriage both changed direction, and both results came after the change: a first sale and 1st place. Evidence: [Codaro](https://shyl.ai/projects/codaro), [mTriage](https://shyl.ai/projects/mtriage) ### I learn a new field in months, not years. OpenRoad’s Android app was my first mobile app, and it was live in three months. AI agents and smart contracts came the same way: learnt for a product that shipped. With AI I never stop, because the tools change every month. Evidence: [OpenRoad](https://shyl.ai/projects/openroad), [TaxCourtBlog](https://shyl.ai/projects/taxcourtblog), [PrivateBond](https://shyl.ai/projects/privatebond) ### CTO while still in secondary school. At Codaro I managed the developers, wrote much of the code, and made and presented most of our pitch decks. In one month, we prepared a hackathon for about 90 people. Evidence: [Codaro](https://shyl.ai/projects/codaro) ## Projects ### OpenRoad **$50,000+ in revenue from 50,000+ drivers.** A drive tracker for car enthusiasts: every drive logged automatically with its route and stats, convoy voice with friends, and drives ready to share. Live on iOS and Android, with more than 50,000 drivers. - Status: Shipped (Since Jan 2026) - Role: Technical co-founder. I built the Android app and work across both platforms. - Platforms: iOS, Android - Built with: Kotlin, Jetpack Compose, Swift, Firebase, Mapbox, Google Play Billing - Milestone (as of Oct 2026): More than 50,000 drivers and more than $50,000 in revenue - Case study: https://shyl.ai/projects/openroad - openroaddrive.com: https://www.openroaddrive.com - App Store: https://apps.apple.com/us/app/open-road-drive-tracker/id6755834573 - Google Play: https://play.google.com/store/apps/details?id=com.mikita.openroad - Instagram: https://www.instagram.com/openroad.app/ - TikTok: https://www.tiktok.com/@openroad.app - X: https://x.com/openroad_app #### What I did - Built the Android app from scratch as the only Android developer, from January to launch on 31 March 2026 - Since launch, develop features on both Android and iOS - Own subscriptions and billing on Android - Help with marketing, which my co-founder leads #### Metrics - 50,000+ drivers - $50,000+ subscription revenue - 4.7 App Store rating Team results. Drivers and rating as shown on openroaddrive.com, as of Oct 2026. #### In short A drive tracker for car enthusiasts, founded by Rasputin Carlson in Seattle. I built the Android app alone, billing included, and now ship on both iOS and Android. Best month above $10K; rated 4.7 on the App Store. #### Overview OpenRoad is one of the two projects I am proudest of, alongside Codaro, and my first project with large-scale traction. My co-founder Rasputin Carlson, from Seattle, founded it and built the iOS app. I joined as technical co-founder in January 2026, two days before the iOS launch, and took on Android. It was not my project from the very beginning, which is why I count Codaro as the first product I built from scratch that reached traction. But OpenRoad is where I first worked on something used by tens of thousands of people. #### Problem and users Runners and cyclists have Strava. Drivers who love their cars and the roads they drive have nothing comparable: no simple way to keep a record of their drives, see their stats, or share a great drive with friends. #### My role and collaborators My official title is technical co-founder; Rasputin is the founder. I was the only person building Android: I started in January and we published the Android app on 31 March 2026. Since then I work on both Android and iOS. Rasputin also writes code and leads marketing: short-form video, track-day filming and work with UGC creators. I help with marketing a little. Results like downloads and revenue belong to the team; the Android app and its billing are my work. #### What we built OpenRoad logs every drive automatically the moment the car moves, with the route, speed and stats. Drivers can talk to friends in convoy voice, drop markers on the map, see insights about their driving, navigate, and share a drive to Instagram with the overlay already built in. The app is available in 30 languages. My part was a native Android app in Kotlin and Jetpack Compose with maps and drive recording, Google Play Billing for the paid tier, and a Firebase backend shared with iOS. It was my first mobile app, and I also work in Swift on the iOS side. #### Engineering decisions Building a second platform for an app that already had users meant matching what iOS users were used to, without copying iOS conventions onto Android. Both apps share one backend, so every data change had to work for both clients. Subscriptions were the part that most needed to be boring. When a purchase completed but the server verification call failed, a paying user could end up without access. I changed the flow so access is granted locally as soon as Google Play confirms the purchase, verification retries in the background, and the server asks Play to resend notifications it cannot match yet instead of discarding them. #### Product and business decisions Growth came mostly from short-form video. Our best month was around April 2026, when two of our reels passed one million views each and revenue passed $10,000 in a single month. Marketing combines our own content, track-day filming and UGC creators, and the community grew to more than 30,000 followers on Instagram and over 1,000 members on Discord. #### Outcomes and evidence Live on the App Store and Google Play, with more than 50,000 drivers, a 4.7 rating on the App Store and 30 languages, as shown on openroaddrive.com. More than $50,000 in revenue from in-app subscriptions, with more than $10,000 in our best month, around April 2026. Two reels with more than one million views each, more than 30,000 Instagram followers and a Discord community of over 1,000 people. #### Lessons and next steps In June 2026 Rasputin and I rented an apartment in Warsaw and lived and worked there together for a month, fully focused on OpenRoad. We shipped many new features in that month. Working that closely with a co-founder from the US also made my English much better. OpenRoad was my first mobile app and my first time on a product used at this scale, with many challenges along the way. The next step is better marketing: the product is there, and I believe it still has a future. #### How it went - Jan 2026: Joining OpenRoad as technical co-founder. I joined two days before Rasputin launched the iOS app, and started building Android. It became my first project with large-scale traction. - Jan 2026: iOS app launches. Built by Rasputin. Traction followed. - 31 Mar 2026: OpenRoad launches on Android. Three months after joining, I published the Android app I built as OpenRoad's only Android developer. Since then I work on both platforms. - Apr 2026: Best month: two reels pass one million views - Jun 2026: A month locked in in Warsaw. Rasputin flew in from Seattle and we rented an apartment in Warsaw for a month, living together and working only on OpenRoad. We shipped many new features, and my English got much better. ### Codaro **Google for Startups residency. Three markets tested, the third one paid.** Developer-insights software that turns live coding activity into a clear, real-time picture of what a software team is doing, without status meetings or micromanagement. My largest project, and the first I built from the start that reached paying users. - Status: Shipped (Mar 2025 – May 2026) - Role: Co-founder and CTO. - Built with: Next.js, TypeScript, Google Cloud Platform, VS Code and IntelliJ plugins, Real-time streaming connectors, AI code-analysis pipelines, LLM summaries, MCP server - Milestone (as of 16 Mar 2026): First paying user after two pivots, on the B2C version - Case study: https://shyl.ai/projects/codaro - codaro.dev: https://codaro.dev #### What I did - Led engineering and managed the development team, while writing a large share of the code myself - Shaped product and company direction together with Radek, our CEO - Created and presented most of our pitch decks, for competitions, juries and investors - Marketing: short-form video and social content - Co-organised the February 2026 hackathon #### In short Developer insights for software teams: IDE plugins, real-time streaming and AI code analysis on Google Cloud. Freelancers first, then engineering teams; individual developers were the market that paid, from March 2026. As CTO I led the developers and wrote much of the code. #### Overview Codaro is my largest project and the one I developed the most. I worked on it for more than a year, from March 2025 to May 2026. Radosław Soysal and I started it on the night train back from the Econverse V4 final, and in practice it was mostly Radek's and my project. It became my first registered company, a Polish sp. z o.o., which still operates. In that time Codaro won two awards at TeenCrunch, spent eight months at Google for Startups Campus, organised a hackathon for about 90 people, changed direction twice, and made its first sale. It was the first product I built from the very beginning that reached real traction. #### Problem and users We started from a problem we had ourselves as freelance developers. Developers are often paid by the hour, and the person paying has no live view of what is actually being done. Reported hours and real work drift apart, some developers inflate their time, and neither side fully trusts the other. After the first pivot, the users became engineering leads. As teams grow and AI coding tools multiply the volume of code, managers lose sight of progress, spend hours in status meetings, and discover blockers and bottlenecks too late. #### Pivot 1: from freelancers to teams (Sept 2025) - Before: Time tracking that proves real hours: a plugin that counts actual coding time and describes with AI what a freelance developer did. - After: A live view for engineering teams: progress, blockers, bottlenecks and code review from real coding activity, instead of status meetings. Why we changed course: - We concluded that almost nobody would pay to verify hours, and those who would, would pay very little. The bigger problem was the flood of code: teams produce more of it than ever, and leads cannot see where work is stuck. #### Pivot 2: a B2C test (2026) - Before: Selling a team product to companies, through investor pitches and B2B conversations. - After: The same plugin offered directly to individual developers: it reviews their code and finds bottlenecks. Why we changed course: - A smaller change that reused what we had already built, to test whether individual developers would pay faster than companies. We ran it for about three months in early 2026, and it worked: the B2C version brought our first sale on 16 March 2026. #### Pivot 3: back to B2B, by the team (2026) _After I left. I was not involved in this change._ - Before: Insights into developer work, for teams and for individual developers. - After: A European control layer for AI coding agents: an agent writes code, an independent auditor reviews it before deployment. See codaro.dev. #### My role and collaborators I was co-founder and CTO, and Radek was CEO. We agreed on that split because I did not want to carry the whole company at the time, but in practice my role reached well beyond engineering. I managed the developers, Radek and Dominik Dyga, wrote a large part of the code myself, made and presented most of our pitch decks, and created marketing content, including short videos. In the summer of 2025 I regularly spent 15-hour days coding. The co-founders were Radek, Dominik and me. The starting team also included Wiktor Dźwigałowski, whom I met with Radek at Econverse V4, and Jan Matecki, Wiktor's friend. Dominik is Radek's friend. Kacper Kamiński joined during the TeenCrunch accelerator, where I was the only participant working on two projects at once. Awards, the residency and the hackathon were team results. #### What we built IDE plugins for VS Code and IntelliJ that capture real coding activity without interrupting the developer, and real-time connectors that stream that activity to the backend and on to a live dashboard. AI pipelines that analyse code changes as they happen: plain-language summaries of what was done, review signals on the code itself, and early warnings about blockers, risky changes and bottlenecks. A web dashboard for team leads with a live feed, team statistics and alerts, and later a B2C version of the plugin for individual developers that reviews their code and points out bottlenecks. #### Engineering decisions The core decision was to measure from the editor rather than the task board. Tickets show what someone said they did; editor activity and commits show what actually happened. The hard part was making that useful without turning it into surveillance, so the product summarises work for the developer's benefit as well, instead of streaming raw activity to a manager. Everything ran on Google Cloud Platform. Because the product sees source code, trust was a technical requirement, not a marketing line: data stayed in the EU, customer code was never used to train models, and data was encrypted in transit and at rest. An MCP server let teams connect their own tools, and the plugins plus Git and Jira integrations could be set up in a few minutes. #### Product and business decisions We made a first deck, pitched it at several events and applied to the TeenCrunch accelerator. Pitches and conversations with managers and developers kept ending in the same place, which led to the first pivot in September 2025. After it we pitched to many investors. The business model was a per-seat subscription with a free trial, aimed at software agencies, startups and larger teams. When B2B sales were slow, we tested the same product on individual developers instead, promoted it, and made our first sale. #### Outcomes and evidence TeenCrunch Demo Day, 13 June 2025: the Audience Award and, as Rzeczpospolita reported, a place at Google for Startups. The same day mTriage, my other project in the accelerator, won first place. Residents at Google for Startups Campus in Warsaw from June 2025 to February 2026, alongside a cohort of startups selected for the programme: a coworking space, training, team retreats and many conversations with other founders. Hackathon Codaro, 24 February 2026, at the Campus: 16 teams, about 90 participants and over 25,000 PLN in prizes. We had exactly one month to arrange the agenda, guests, participants, funding, venue, sponsors, partners and marketing. As far as I know, we were the only startup from the programme to organise its own event there. The team organised a second hackathon in July 2026. First paying user on 16 March 2026, from the B2C version: real traction for a product I had built from the very beginning. #### Lessons and next steps Talking to the people who would pay mattered more than how well we built. Two pivots in a year felt slow at the time, but each came from hearing the same objection again and again, and the second one is what finally produced a sale. From January 2026 I was also building OpenRoad's Android app, and in May I stepped back from Codaro, for my own reasons, to focus on it. I am still an advisor to the team and good friends with all the founders. Radek continues as CEO with Kacper and Jan, and has since taken Codaro in a new direction. #### How it went - Mar 2025: Codaro starts on a night train. Travelling back to Warsaw from the lost Econverse V4 final in Łódź, Radosław Soysal and I decided to start Codaro together. - Apr 2025: Into the TeenCrunch accelerator; Kacper Kamiński joins. After pitching the first deck at several events. - 13 Jun 2025: Two projects, one Demo Day. I worked on two projects in the TeenCrunch accelerator at the same time. At the Demo Day, mTriage won first place, and Codaro won the Audience Award and a place at Google for Startups. - Jun 2025: Google for Startups Campus residency. Codaro moved into Google for Startups Campus in Warsaw, the place it won at TeenCrunch. Coworking, training and a cohort of other founders until February 2026. - Sept 2025: Codaro's first pivot: from freelancers to teams. Few would pay to verify a developer's hours. We moved from time tracking to a live view of progress and blockers for whole teams. - 2026: Codaro's second pivot: testing B2C. For about three months in early 2026 we offered the same plugin directly to individual developers, to see if they would pay faster than companies. It brought our first sale in March 2026. - 24 Feb 2026: Hackathon Codaro. A hackathon at Google for Startups Campus in Warsaw: 16 teams, about 90 participants and over 25,000 PLN in prizes, organised by the Codaro team in one month. - 16 Mar 2026: Codaro's first paying user. The B2C version of Codaro made its first sale: the first traction for a product I had built from the very beginning. - May 2026: Stepping back from Codaro. After more than a year, I stepped back from Codaro to focus on OpenRoad. I stay on as an advisor, and Radek continues as CEO. - 2026: The team pivots to a control layer for AI coding agents. After I stepped back. - Jul 2026: Second Codaro hackathon. Organised by the team. #### Recognition - TeenCrunch Audience Award (Fundacja Youth Business Poland, 13 Jun 2025): Audience Award, with Codaro. https://shyl.ai/awards#teencrunch-codaro - Google for Startups Campus residency (Google for Startups Campus Warsaw, Jun 2025 – Feb 2026): Resident, with Codaro. https://shyl.ai/awards#google-for-startups-residency #### Press - 25 Aug 2026, My Company Polska: [Co jeśli sztuczna inteligencja się pomyli? Polski startup znalazł sposób na kontrolę nad AI](https://mycompanypolska.pl/artykul/co-jesli-sztuczna-inteligencja-sie-pomyli-polski-startup-znalazl-sposob-na-kontrole-nad-ai/21078). A feature on Codaro's current direction as a control layer for AI coding agents, published after I stepped back. - 12 Feb 2026, Gazeta Prawna: [3 godziny kodowania, pizza i 25 tys. PLN w puli. Rusza Codaro Coding Challenge w Google Campus](https://www.gazetaprawna.pl/nowe-technologie/artykuly/10953146,3-godziny-kodowania-pizza-i-25-tys-pln-w-puli-rusza-codaro-coding-c.html). Coverage of the Codaro Coding Challenge ahead of the event. - 11 Feb 2026, Android.com.pl: [Pierwszy w Polsce hackathon symulujący realia pracy w nowoczesnym startupie startuje w Google for Startups Campus](https://android.com.pl/tech/1017191-pierwszy-w-polsce-hackathon-symulujacy-realia-pracy/). Announcement of the Codaro Coding Challenge, the hackathon our team organised. ### TaxCourtBlog A fully automated legal news site: a pipeline of five AI agents turns US Tax Court PDF decisions into published articles, with no human in the loop. - Status: Shipped (Jan 2026 – Feb 2026) - Role: Built it end to end, for David Brunk. - Built with: Next.js, TypeScript, AI agents, Google Cloud - Case study: https://shyl.ai/projects/taxcourtblog - taxcourtblog.com: https://www.taxcourtblog.com #### What I did - Designed the five-agent pipeline from PDF to published post - Built the SEO-focused Next.js site and its deployment #### Overview TaxCourtBlog publishes news about US tax law without anyone writing it by hand. A pipeline of five agents collects new decisions published as PDFs on US court websites, reads them, and turns them into articles that are posted to the site automatically. #### What we built The publishing pipeline, plus a fast, search-friendly site with topic guides, a news feed of recent decisions, site search and an RSS feed. ### Cited A CRM for real estate agencies with a knowledge base: track and close sales more easily, and ask questions about your documents with answers that always cite their source. - Status: In progress (Jun 2026 – Aug 2026) - Role: Built the product. - Built with: Python, FastAPI, React, Gemini, Google Cloud - Case study: https://shyl.ai/projects/cited - cited.pl: https://cited.pl #### What I did - Built the document Q&A engine (RAG) and its citation checks - Built the backend and chat interface #### Overview Cited helps real estate agencies keep track of sales and the documents behind them. Its knowledge base answers questions about contracts, permits and other technical documents. The product is built and waiting for its next step: first clients, partners or investors. #### Engineering decisions The key rule is that it must not make things up. Every factual statement in an answer carries a quote from a document, and the backend checks in code that the quote really appears in the source. If it does not, the answer says the information is not in the documents. ### Kliniqa A voice receptionist for dental clinics, built on ElevenLabs, with a CRM made for dentistry: calendar, patient records and automated messages. - Status: Paused (May 2026 – Jul 2026) - Role: Built the product, with some help from Radosław Soysal. - Built with: ElevenLabs, React, TypeScript, Vite, Firebase - Case study: https://shyl.ai/projects/kliniqa - kliniqa.pl: https://kliniqa.pl #### What I did - Built the voice agent that answers the clinic's phone - Built the reception dashboard: calendar, patients, SMS reminders, team and settings #### Overview Dental clinic receptions spend much of the day on the phone. Kliniqa adds an AI voice agent that answers calls at the front desk, and a reception dashboard built specifically for dental clinics. The project is paused: neither of us has had the time to take it further. ### mTriage **My first startup took 1st place at TeenCrunch.** Software for paramedics at mass-casualty incidents: record patients quickly, coordinate teams and hand over clean documentation. It started as AI triage for hospital emergency departments. - Status: Archived (Oct 2024 – Jul 2025) - Role: Co-founder. I built the software. - Platforms: Web - Built with: Next.js, Supabase, WebSockets, Python and the OpenAI API (first version) - Milestone (as of 13 Jun 2025): 1st place at TeenCrunch - Case study: https://shyl.ai/projects/mtriage - Live demo (sample data): https://mtriage-mvp.vercel.app - Rzeczpospolita on the TeenCrunch results: https://www.rp.pl/prawo-dla-ciebie/art42558001-mlodzi-innowatorzy-i-liderzy-biznesu-nagrodzeni-w-ramach-futurecrunch #### What I did - Built the software: the Python prototype, then the Next.js app with live updates - Pitched it at the Econverse final, Warsaw Startup Club and the TeenCrunch Demo Day - Took part in the meetings with the county and the hospital that led to the pivot #### In short Software for paramedics at mass-casualty incidents. It began as AI triage for emergency departments; after meeting the head of Jarocin County and the county hospital’s emergency staff, we rebuilt it for paramedics. I built the software: a real-time app that shows every team live. The original idea was Mikołaj Mikołajczak’s. #### Overview mTriage was my first startup, and the Econverse Cup in Poznań in October 2024 was my first business and startup event. There I met Mikołaj Mikołajczak, who, surprisingly, turned out to be from my own town, Jarocin. He came up with the idea of AI triage for emergency departments, and we started building it together. We reached the Econverse final in December 2024 but did not win. We wanted to keep going, and over the following months the idea changed from AI triage in hospitals into working software for paramedics at mass-casualty incidents, which won TeenCrunch in June 2025. #### Problem and users The first version targeted hospital emergency departments, where patients can wait a long time before anyone assesses how urgent their case is. We wanted patients to answer a short questionnaire on a tablet and have AI suggest a triage priority. After the pivot, the users were paramedic teams. At a mass-casualty incident, such as a serious road or tram accident, they record each patient on long paper forms, by hand and often wearing gloves. The result is slow to fill in, hard to read and hard to pass on to the next team or the hospital. #### The pivot (Apr 2025) - Before: AI triage for hospital emergency departments. Patients answer questions on a tablet and AI suggests how urgent their case is. - After: Coordination and documentation for paramedics at mass-casualty incidents. Why we changed course: - We met the head of Jarocin County, and then organized a meeting with the staff of the county hospital's emergency department. Together with a paramedic who joined as our advisor, that pointed us to a sharper problem: documentation and coordination when many patients need help at once. #### My role and collaborators I co-founded mTriage and built the software. Mikołaj came up with the original idea. Dominik Stojczyk, a paramedic, advised us and tested the app. The awards, meetings and pitches were team results. #### What we built The first version, shown at the Econverse final, ran on a tablet in the waiting room. The patient described the problem in their own words, AI generated follow-up questions with tap-to-answer options, and the triage nurse received a summary with a suggested triage category to approve. A web app for paramedic teams. Each patient is recorded in a few steps (injuries, consciousness, breathing, medication and dose), and patients are grouped by triage colour. The team leader sees the status of every team and can send a patient to a chosen hospital. Instead of a handwritten card, the app produces a clean, printable dispatch card that follows the structure of the standard paper form. #### Engineering decisions The first prototype, the emergency-department questionnaire, was written in Python. For the mass-casualty version I rebuilt the app in Next.js with a Supabase database. At a mass-casualty incident several teams record patients at the same time, and the team leader decides where each one goes. Updates therefore travel over WebSockets: a patient added by one team appears on the leader's screen straight away, without anyone refreshing. The app shows its live connection status and latency, because teams in the field cannot assume a stable signal. #### Product and business decisions We planned to license the software per hospital: a free tier with limited features, an annual licence and a one-time licence, each including tablet integration, support and training for paramedics. Hospitals and counties buy through public digitalisation programmes, so we mapped which funding programmes could pay for it. We also considered selling through established hospital-software vendors instead of only directly. #### Outcomes and evidence First place (5,000 PLN) at the TeenCrunch Demo Day on 13 June 2025, chosen by the jury from the accelerator's teams. mTriage also received media support from Rzeczpospolita, which covered the results. Working software, tested with paramedic teams. Our paramedic advisor presented it at one of the largest training events for paramedics in Poland. A paramedic who tested it said: “This is literally what I wanted to have.” #### Lessons and next steps The most useful step was letting go of the first idea. Meeting the people who run emergency care in our county showed us a problem they actually wanted solved. In July 2025 we froze mTriage, and later archived it, because I was focusing on Codaro. The next steps we had planned were registering a company, certifying the app as a Class I medical device under the EU MDR, field tests, and integration with ambulance teams' tablets. #### How it went - Oct 2024: My first startup event. At the Econverse Cup in Poznań I met Mikołaj Mikołajczak, who turned out to be from my town, Jarocin. We started mTriage there and reached the final in December 2024. - Nov 2024: Meetings with the Jarocin County head and the county hospital's emergency department. The first conversations with the people who run emergency care in our county. - Dec 2024: Econverse Cup final. We reached the final but did not win. - Apr 2025: mTriage pivots to mass-casualty incidents. As we joined the TeenCrunch accelerator, and after meeting the Jarocin County head and the county hospital's emergency department, we moved from AI triage in hospitals to software for paramedics at mass-casualty incidents. - 13 Jun 2025: First place at TeenCrunch. At the Demo Day closing the TeenCrunch accelerator, the jury gave mTriage first place. Rzeczpospolita covered the results. - Jul 2025: Freezing mTriage. A month after winning TeenCrunch, we froze mTriage and later archived it. My time went to Codaro. #### Recognition - TeenCrunch (Fundacja Youth Business Poland, 13 Jun 2025): 1st place (5,000 PLN), with mTriage. https://shyl.ai/awards#teencrunch - Econverse Cup (Econverse, Dec 2024): Finalist, with mTriage. https://shyl.ai/awards#econverse-cup #### Press - 23 Jun 2025, Rzeczpospolita: [Młodzi innowatorzy i liderzy biznesu nagrodzeni w ramach FutureCrunch](https://www.rp.pl/prawo-dla-ciebie/art42558001-mlodzi-innowatorzy-i-liderzy-biznesu-nagrodzeni-w-ramach-futurecrunch). TeenCrunch results: first place for mTriage, named with me as author, and a place at Google for Startups for Codaro. ### Riddep A competition concept: AI agents that give small restaurants the kind of data large chains have, from where to open to how to staff and when to order. - Status: Archived (Nov 2024 – Mar 2025) - Role: Product lead in a team of five. I ran the pitches and made most of the mock-ups. - Case study: https://shyl.ai/projects/riddep - riddep.com (an older site, describing an earlier idea): https://riddep.com #### What I did - Led the product - Gave every pitch, including at the Warsaw stage and the final - Made most of the presentations and interface mock-ups #### Overview Riddep was the project my team built for Econverse V4. I applied right after the Econverse Cup in Poznań, and we worked on it from November 2024 to the final at the end of March 2025. It was never meant to become a company. We built it for the competition, using the name and branding of an earlier idea of Radek's. #### Problem and users Small restaurants and cafés compete with chains that have analysts, expensive tools and real-time data. Owners often decide where to open, what to charge and how many people to schedule on instinct. #### My role and collaborators I led the product, gave every pitch and made most of the presentation and interface mock-ups. The team of five included Radosław Soysal and Wiktor Dźwigałowski. #### What we built A concept and interface mock-ups, not a working product. A main AI agent would coordinate specialised agents for competitors, prices, accounting, staffing, feedback and marketing. The first phase helps choose a location using competition, foot traffic and rent; the second suggests staffing by peak hours, delivery schedules and answers questions in a chat. #### Product and business decisions We pitched a freemium subscription: a free dashboard with an AI assistant, and paid tiers adding optimisation, marketing and growth tools. Gastronomy was the entry market, with hotels and other local services as the next step. #### Outcomes and evidence We reached the Econverse V4 final at the end of March 2025. Day one was in Warsaw, at Warsaw Tower, and day two in Łódź. We did not win. #### Lessons and next steps The project itself was not strong, but the people were. Through Riddep I met Radosław Soysal, my co-founder ever since, and Wiktor Dźwigałowski, who told me about the economics olympiad. Both are still close friends. On the night train back from the lost final, Radek and I started Codaro. #### How it went - Mar 2025: Econverse V4 final with Riddep. Two days of finals, first at Warsaw Tower and then in Łódź. We did not win, but through the team I met Radosław Soysal and Wiktor Dźwigałowski, who told me about the economics olympiad. - Mar 2025: Codaro starts on a night train. Travelling back to Warsaw from the lost Econverse V4 final in Łódź, Radosław Soysal and I decided to start Codaro together. #### Recognition - Econverse V4 (Econverse, Mar 2025): Finalist, with Riddep. https://shyl.ai/awards#econverse-v4 ### PrivateBond Secure media sharing for creators: every viewer gets a uniquely, invisibly watermarked copy, so a leak can be traced to the person who shared it. I was the technical person behind its blockchain layer. - Status: Archived (Apr 2025 – Jul 2025) - Role: Freelance developer: smart contracts and blockchain integration. - Built with: Solidity, Hardhat, Next.js, TypeScript, Python - Case study: https://shyl.ai/projects/privatebond - privatebond.io: https://privatebond.io #### What I did - Wrote all of the smart contracts and the blockchain integration - Contributed a little to the watermarking, which was built by others on the project #### Overview PrivateBond is a friend's project. It was my first large freelance project: partly paid work, and partly a lot of unpaid time spent helping a friend. I was not a co-founder or a user of the product. I was the technical person behind its blockchain side, from April to July 2025, with June as the most intensive month. #### Problem and users Creators who share private photos and videos have no way to know who leaked them. PrivateBond gives every viewer a unique, invisibly marked copy, so a leaked file points back to the recipient it came from. #### What we built All of the smart contracts for the version of PrivateBond built on blockchain: an upgradeable core contract, a marketplace with escrowed bids, and integrations that let deposited funds earn yield. Plus the scripts, tests and web integration around them. #### Engineering decisions Smart contracts are unforgiving: once deployed they hold real money and cannot simply be patched. That meant upgradeable contracts behind a proxy, extensive testing on a test network before anything went live, and treating every external call as a security risk. It was the hardest engineering I had done at that point. #### Lessons and next steps I do not count PrivateBond among my own projects, but it was a real technical challenge and my first experience of doing paid work for someone else's product. I stopped working on it in July 2025. The project continues without me, and today's site presents it as forensic watermarking for media sharing. ## Recognition ### XXXIX Olimpiada Wiedzy Ekonomicznej **9th in Poland, out of 5,551 entrants.** - Organizer: Polskie Towarzystwo Ekonomiczne - Date: 9 Apr 2026 - Type: Competition result - Result: Laureate, 9th place in Poland - Stage: 5,551 entrants · 3rd of 394 in stage II · 9th in the final, 8–9 April 2026 - Record: https://shyl.ai/awards#owe-xxxix Poland's national economics olympiad, run by the Polish Economic Society. I became a laureate on my first attempt, nine months after starting economics from zero. My essays scored 25/25 in stage II and 24/25 in the final, the second-highest among 100 finalists. Laureates are admitted to SGH Warsaw School of Economics. I started economics in July 2025 with no background. In April 2026 I finished 9th in the national final and became one of 30 laureates, after a perfect essay score in stage II. Laureates are admitted to SGH Warsaw School of Economics. Sources: - [Edition summary (PTE)](https://owe.pte.pl/informacje/poprzednie-edycje/podsumowanie-xxxix.html) - [Laureates](https://owe.pte.pl/upload/files/uczestnicy/laureaci/laureacixxxixowe.pdf) - [Stage II results](https://owe.pte.pl/upload/files/uczestnicy/zakwalifikowani/xxxix/zakwalifikowanizawodycentralnexxxix.pdf) - [Final written results](https://owe.pte.pl/upload/files/protokoly/protokol-z-czesci-pisemnejetap-centralny-owe-xxxix.pdf) ### Stypendium Prezesa Rady Ministrów - Organizer: Prezes Rady Ministrów - Date: 2026 – 2027 - Type: Scholarship - Result: Prime Minister's scholarship, 2026/2027 school year (second time) - Record: https://shyl.ai/awards#pm-scholarship-2026 A Polish state scholarship awarded by the Prime Minister to secondary-school students for outstanding results. ### TeenCrunch - Organizer: Fundacja Youth Business Poland - Date: 13 Jun 2025 - Type: Competition result - Result: 1st place (5,000 PLN), with mTriage - Stage: Demo Day at the FutureCrunch conference - Record: https://shyl.ai/awards#teencrunch An accelerator for youth startup teams, run by the Youth Business Poland Foundation. The jury chose three winning teams; we came first. mTriage also got media support from Rzeczpospolita. Sources: - [Rzeczpospolita, 23 June 2025](https://www.rp.pl/prawo-dla-ciebie/art42558001-mlodzi-innowatorzy-i-liderzy-biznesu-nagrodzeni-w-ramach-futurecrunch) ### TeenCrunch Audience Award - Organizer: Fundacja Youth Business Poland - Date: 13 Jun 2025 - Type: Competition result - Result: Audience Award, with Codaro - Stage: Demo Day at the FutureCrunch conference - Record: https://shyl.ai/awards#teencrunch-codaro Chosen by the audience at the TeenCrunch Demo Day. At the same event Codaro also won a place at Google for Startups. Sources: - [Rzeczpospolita, 23 June 2025](https://www.rp.pl/prawo-dla-ciebie/art42558001-mlodzi-innowatorzy-i-liderzy-biznesu-nagrodzeni-w-ramach-futurecrunch) ### Google for Startups Campus residency - Organizer: Google for Startups Campus Warsaw - Date: Jun 2025 – Feb 2026 - Type: Residency - Result: Resident, with Codaro - Record: https://shyl.ai/awards#google-for-startups-residency Codaro won its place at the TeenCrunch Demo Day. Eight months at the Campus in Warsaw: a desk, training and a cohort of other founders. Sources: - [Rzeczpospolita, 23 June 2025](https://www.rp.pl/prawo-dla-ciebie/art42558001-mlodzi-innowatorzy-i-liderzy-biznesu-nagrodzeni-w-ramach-futurecrunch) ### Econverse V4 - Organizer: Econverse - Date: Mar 2025 - Type: Finalist - Result: Finalist, with Riddep - Stage: Final in Warsaw and Łódź - Record: https://shyl.ai/awards#econverse-v4 ### Stypendium Prezesa Rady Ministrów - Organizer: Prezes Rady Ministrów - Date: 2025 – 2026 - Type: Scholarship - Result: Prime Minister's scholarship, 2025/2026 school year - Record: https://shyl.ai/awards#pm-scholarship-2025 A Polish state scholarship awarded by the Prime Minister to secondary-school students for outstanding results. ### Econverse Cup - Organizer: Econverse - Date: Dec 2024 - Type: Finalist - Result: Finalist, with mTriage - Record: https://shyl.ai/awards#econverse-cup ## Journey (oldest first) ### 2008: Where it started: Minsk I was born in Minsk, the capital of Belarus. Looking back, the business side showed up early: as a child I collected stones on the street, opened a shop in our flat and sold them to my family. I do not even remember what I was saving for. I just wanted to earn my own money, and I kept saving from then on. Permalink: https://shyl.ai/journey#2008-minsk ### 2019: First programming classes at IT Club At eleven I started programming classes at IT Club, because I wanted to solve technical problems and build things of my own. Over four years I went from C to HTML and CSS, then C#, then Unity. One language a year: C first, then HTML and CSS the next year, then C#, then Unity. That is where my interest in building things for real started. Permalink: https://shyl.ai/journey#2019-it-club ### Apr 2022: Moving to Poland In April 2022 I moved to Poland. Permalink: https://shyl.ai/journey#2022-poland ### 2023: Teaching myself Python In the summer of 2023 I bought Python courses on Udemy with my own money and worked through them on my own. It was the first time I chose what to learn next by myself. Permalink: https://shyl.ai/journey#2023-python ### Oct 2024: My first startup event At the Econverse Cup in Poznań I met Mikołaj Mikołajczak, who turned out to be from my town, Jarocin. We started mTriage there and reached the final in December 2024. Permalink: https://shyl.ai/journey#2024-econverse-start ### Nov 2024: Joining Econverse V4 Right after the Econverse Cup I applied to Econverse V4 and joined a team of five to build Riddep. Permalink: https://shyl.ai/journey#2024-econverse-v4-start ### Mar 2025: Econverse V4 final with Riddep Two days of finals, first at Warsaw Tower and then in Łódź. We did not win, but through the team I met Radosław Soysal and Wiktor Dźwigałowski, who told me about the economics olympiad. Permalink: https://shyl.ai/journey#2025-econverse-v4-final ### Mar 2025: Codaro starts on a night train Travelling back to Warsaw from the lost Econverse V4 final in Łódź, Radosław Soysal and I decided to start Codaro together. Permalink: https://shyl.ai/journey#2025-codaro-founded ### Apr 2025: mTriage pivots to mass-casualty incidents As we joined the TeenCrunch accelerator, and after meeting the Jarocin County head and the county hospital's emergency department, we moved from AI triage in hospitals to software for paramedics at mass-casualty incidents. Permalink: https://shyl.ai/journey#2025-mtriage-pivot ### 13 Jun 2025: First place at TeenCrunch At the Demo Day closing the TeenCrunch accelerator, the jury gave mTriage first place. Rzeczpospolita covered the results. Evidence: - [Rzeczpospolita](https://www.rp.pl/prawo-dla-ciebie/art42558001-mlodzi-innowatorzy-i-liderzy-biznesu-nagrodzeni-w-ramach-futurecrunch) Permalink: https://shyl.ai/journey#2025-teencrunch ### 13 Jun 2025: Two projects, one Demo Day I worked on two projects in the TeenCrunch accelerator at the same time. At the Demo Day, mTriage won first place, and Codaro won the Audience Award and a place at Google for Startups. Evidence: - [Rzeczpospolita](https://www.rp.pl/prawo-dla-ciebie/art42558001-mlodzi-innowatorzy-i-liderzy-biznesu-nagrodzeni-w-ramach-futurecrunch) Permalink: https://shyl.ai/journey#2025-teencrunch-two-projects ### Jun 2025: Google for Startups Campus residency Codaro moved into Google for Startups Campus in Warsaw, the place it won at TeenCrunch. Coworking, training and a cohort of other founders until February 2026. Permalink: https://shyl.ai/journey#2025-gfs-residency ### Jul 2025: Freezing mTriage A month after winning TeenCrunch, we froze mTriage and later archived it. My time went to Codaro. Permalink: https://shyl.ai/journey#2025-mtriage-ends ### 1 Jul 2025: Starting economics from zero Wiktor Dźwigałowski told me about the economics olympiad, and on 1 July 2025 I started learning economics with no background at all. I remember not understanding why price and quantity sit on the axes the way they do in a supply-and-demand chart. Permalink: https://shyl.ai/journey#2025-economics-start ### Sept 2025: Codaro's first pivot: from freelancers to teams Few would pay to verify a developer's hours. We moved from time tracking to a live view of progress and blockers for whole teams. Permalink: https://shyl.ai/journey#2025-codaro-pivot ### 2025: Prime Minister's scholarship Awarded the Prime Minister's scholarship for the 2025/2026 school year, and again for 2026/2027. Permalink: https://shyl.ai/journey#2025-pm-scholarship ### Jan 2026: Joining OpenRoad as technical co-founder I joined two days before Rasputin launched the iOS app, and started building Android. It became my first project with large-scale traction. Permalink: https://shyl.ai/journey#2026-openroad-join ### 2026: Codaro's second pivot: testing B2C For about three months in early 2026 we offered the same plugin directly to individual developers, to see if they would pay faster than companies. It brought our first sale in March 2026. Permalink: https://shyl.ai/journey#2026-codaro-b2c ### 24 Feb 2026: Hackathon Codaro A hackathon at Google for Startups Campus in Warsaw: 16 teams, about 90 participants and over 25,000 PLN in prizes, organised by the Codaro team in one month. We had exactly one month for everything: the agenda, guests, participants, funding, sponsors, partners and marketing. It taught me more about running something under a hard deadline than any sprint. Permalink: https://shyl.ai/journey#2026-hackathon ### 16 Mar 2026: Codaro's first paying user The B2C version of Codaro made its first sale: the first traction for a product I had built from the very beginning. Permalink: https://shyl.ai/journey#2026-codaro-first-sale ### 31 Mar 2026: OpenRoad launches on Android Three months after joining, I published the Android app I built as OpenRoad's only Android developer. Since then I work on both platforms. Permalink: https://shyl.ai/journey#2026-openroad-android ### 9 Apr 2026: Laureate of the national economics olympiad, first attempt 9th place in Poland in the central final of the XXXIX Olimpiada Wiedzy Ekonomicznej, after placing 3rd in the country out of 394 participants in stage II. My goal had been to reach the final; I came out a laureate, with admission to SGH. The essay was what I feared most. Polish is my third language, and four years earlier I could not speak it. The task is to write seven or more pages on an economics topic you only see on the day. I practised until it became my strength: a maximum 25/25 in stage II and 24/25 in the central final, the second-highest essay score among the 100 finalists. Most people need two or three attempts to become a laureate. I studied every day for nine months, covered at least the first year of an economics degree, and did it alongside school and my projects. Evidence: - [Official list of laureates](https://owe.pte.pl/upload/files/uczestnicy/laureaci/laureacixxxixowe.pdf) Permalink: https://shyl.ai/journey#2026-owe-final ### May 2026: Stepping back from Codaro After more than a year, I stepped back from Codaro to focus on OpenRoad. I stay on as an advisor, and Radek continues as CEO. Permalink: https://shyl.ai/journey#2026-codaro-step-back ### Jun 2026: A month locked in in Warsaw Rasputin flew in from Seattle and we rented an apartment in Warsaw for a month, living together and working only on OpenRoad. We shipped many new features, and my English got much better. Permalink: https://shyl.ai/journey#2026-openroad-warsaw ## Press - 25 Aug 2026, My Company Polska: [Co jeśli sztuczna inteligencja się pomyli? Polski startup znalazł sposób na kontrolę nad AI](https://mycompanypolska.pl/artykul/co-jesli-sztuczna-inteligencja-sie-pomyli-polski-startup-znalazl-sposob-na-kontrole-nad-ai/21078). A feature on Codaro's current direction as a control layer for AI coding agents, published after I stepped back. (Does not name Mikita.) - 12 Feb 2026, Gazeta Prawna: [3 godziny kodowania, pizza i 25 tys. PLN w puli. Rusza Codaro Coding Challenge w Google Campus](https://www.gazetaprawna.pl/nowe-technologie/artykuly/10953146,3-godziny-kodowania-pizza-i-25-tys-pln-w-puli-rusza-codaro-coding-c.html). Coverage of the Codaro Coding Challenge ahead of the event. (Does not name Mikita.) - 11 Feb 2026, Android.com.pl: [Pierwszy w Polsce hackathon symulujący realia pracy w nowoczesnym startupie startuje w Google for Startups Campus](https://android.com.pl/tech/1017191-pierwszy-w-polsce-hackathon-symulujacy-realia-pracy/). Announcement of the Codaro Coding Challenge, the hackathon our team organised. (Does not name Mikita.) - 23 Jun 2025, Rzeczpospolita: [Młodzi innowatorzy i liderzy biznesu nagrodzeni w ramach FutureCrunch](https://www.rp.pl/prawo-dla-ciebie/art42558001-mlodzi-innowatorzy-i-liderzy-biznesu-nagrodzeni-w-ramach-futurecrunch). TeenCrunch results: first place for mTriage, named with me as author, and a place at Google for Startups for Codaro. ## Recommendations - Bruno Machaj, Program Manager, Google for Startups (3 Apr 2026). Ran the Google for Startups Campus Warsaw residency, where I was the youngest resident in the program's history (June 2025 to February 2026). Quote (translated from Polish): "It is rare to meet someone at the start of their career who combines technical skill, business thinking and people skills so naturally. The recommendation is simple: this is a talent worth betting on." Full letter: https://shyl.ai/recommendations/machaj-google-for-startups-2026.pdf - David N. Brunk, Attorney at Law (2 Apr 2026). A US client who hired me for software projects for more than a year. Quote: "Mikita is the kind of person who rewards investment. He is honest, capable, and motivated. I recommend him without reservation." Full letter: https://shyl.ai/recommendations/brunk-2026.pdf ## Collaboration New startups, partnerships and products worth building. I’m always up for a good conversation. - Co-founding early-stage startups - Building products with founders and partners - Business partnerships around AI, mobile and cloud products