I'm Tomasz Kapcia. I take software end-to-end: sit down with the people who have the problem, name it, build the fix and stay until it runs in production. Engineer at Codility by day, co-founder of Two Bit Coders after hours; an MSc in Computer Science with an AI thesis sits behind both. Founded my first software company at 18.
Full-stack engineer in Product Engineering, working across Screen, Interviews and Skills Intelligence. I helped take Skills Intelligence to its first enterprise customers; in its debut G2 season it was named a Leader in Skills Management. Promoted after my first year.
Co-founder of a boutique two-person software studio: AI enablement and software that digitizes operations. Owners bring scattered tools and manual routines; they get one system and their time back.
Low-level programming for the love of the game: a hobby Python compiler written in Rust. It keeps me honest about how languages actually work.
Toolbox React and TypeScript, Next.js, Expo, Python and Django. ML: TensorFlow, scikit-learn, XGBoost, SHAP. AI daily drivers: Claude Code, ChatGPT, Ollama, Automatic1111, OpenAI and Anthropic APIs. After hours: Rust, C, LLVM.
Most of my client work is what the industry now calls forward-deployed engineering: embedded with the owner, shipping automations that win back their time.
built on Claude Code and MCP servers, deployed inside a client's business, so the owner's know-how stops living only in the owner's head.
code review culture, guardrails and playbooks that make AI output dependable enough to merge. At Codility, I was part of the team whose standards unified output quality and cut review times.
for developers and founders, including the dev team of an ERP software company. Practice runs on the team's own real work.
grounded in an MSc thesis on ML detection of emotion-recognition deficits: XGBoost, SHAP, explainability. Graded 6/6, applied in a university research project.
A new product line built on the company's existing foundations: my team owns it end-to-end, from research to live customers. One piece began as an internal-hackathon experiment, AI-generated candidate feedback that held up and is now permanent, read by candidates every day.
Led the build of an AI-backed app that helps diagnose and train emotion recognition in children. Born at a 24-hour hackathon (2nd place), grown into interdisciplinary university research, presented at the Silesian Science Festival.
A hobby compiler, not an interpreter: it takes CPython bytecode straight to LLVM, skipping the AST. So far it compiles integer arithmetic, variables and print. Built purely for the joy of it. github.com/Tomek20225
Reverse-recruitment platform: one codebase for Android, iOS and web (Expo + Next.js). Built as one of the first 20 Polish participants in Universities of the Future, the pilot program of NCBR, Poland's National Centre for Research and Development, pairing students with industry experts. National finalist of the Young Masters competition at the 29th Teleinformatics Forum.
Architected a service for 70,000 users, with contests running live for up to 10,000 participants at once.
This page is the canonical version of my CV. Expand any role for the full story.
Full-stack work across Python, Django, React and TypeScript. Since 2026 I have been in Product Engineering, working across Screen, Interviews and Skills Intelligence. Right now that means AI for Business, Codility's line for assessing how people in business roles work with AI: I work on its task editor and helped the team of psychologists calibrate its deterministic scoring. Before that I was on the team that carried Skills Intelligence, the module that turns skills data into decisions leaders can trust, from concept and research all the way to live enterprise customers. In its first season in the G2 Skills Management category (Summer 2026), it was named a Leader and a High Performer across the board, with enterprise recognition for best results, best estimated ROI, and fastest and best implementation. Co-built the AI-generated feedback that candidates receive after their sessions, and was part of the team that set engineering standards which unified output quality across the org and cut review times. Fully remote, promoted to IC4 after the first year. Recruitment technology keeps finding me: I hired for my own software house, built a reverse-recruitment platform with NCBR, and now I build the tools other teams hire and grow with.
A two-person studio run with one principle: the people who sell the work are the people who build it. The work is forward-deployed by nature: we sit down with the owner, run the discovery, then build the fix and stay through the production rollout. The practice covers AI enablement built on Claude Code (audits, hands-on training, engineering standards) that ends with working automations in the owner's hands, and that line is mine end-to-end; digitizing company operations into single dedicated systems, like the one we are rolling out for a road transport company, replacing hand-run spreadsheets and pulling scattered tools into one place; and web presence that earns client trust, WCAG accessibility included. Boutique by design, with a deliberately short client queue.
Software house specialized in custom web and hybrid apps. Grew it from a teenage idea to a team of 10 with over 1M PLN in annual revenue; oversaw strategy, sales and delivery while staying hands-on with code. The company shipped 100+ projects, including an education platform architected for 70,000 users. Mentored 15 interns into working developers. Handed the company to my partner in 2025, took the lessons with me.
Pixel-perfect client websites built from design files, WordPress integrations, requirements gathering and technical specs. First taste of production responsibility.
AI thesis graded 6/6: ML detection of emotion-recognition deficits (XGBoost, SHAP).
Specialized in web application design.
Deep learning, big data and data mining.
The interesting problems always arrive by email. Client work runs through Two Bit Coders; everything else lands here.
me [at] tomasz-kapcia.com
For recruiters: I consent to the processing of the personal data published on this page for recruitment purposes, in line with Regulation (EU) 2016/679 (GDPR), where applicable. You are welcome to contact me directly; no prior arrangement is needed.