Software Engineer · Poland
I'm a software engineer focused on building full-stack products and AI-powered tools. Currently working on fraud detection systems for a large global marketplace. Before that, a software house where I shipped across a wide range of products and domains. I care about code that holds up under load, systems that stay legible as they grow, and AI features that earn their place in production.
Maintaining and extending the fraud detection systems that protect a $70B+ GMV marketplace operating across 220+ countries. The pipeline processes 50–100M events per day at sub-100ms latency, with detection logic spanning 12–15 core services and tight integration with ML models. Work closely with data science and operations to ship detection rules, close abuse vectors, and keep the platform trustworthy at scale.
Worked on the tooling and migration infrastructure behind a monorepo transition from Gradle to Bazel — a 4,000+ project effort used by 1,000+ engineers. Built migration tooling and provided hands-on support that helped teams move to the new build system without disrupting active development. The migration cut local build times 3–5x, halved CI times, and lifted developer satisfaction by 26 percentage points.
Led the design and delivery of an internal reservation platform adopted across the company, replacing a patchwork of manual workflows for ~400 employees. Owned the full stack: REST API, React frontend, cloud infrastructure on Azure, and the CI/CD pipeline that shipped it. Drove the architecture, scoped the rollout with stakeholders, and kept the system reliable as adoption grew.
Designed and built websites for local businesses — handling everything from initial requirements to deployment. Developed a practical understanding of the full delivery cycle: working directly with clients, managing scope, and shipping things that worked reliably without a team behind you.
Thesis Research on applying state-of-the-art AI to modern healthcare — exploring how LLMs and retrieval systems can give clinicians trustworthy, evidence-backed answers grounded in the latest scientific literature. Built and evaluated an end-to-end platform that turns hours of manual research into minutes, with a focus on answer quality, source traceability, and clinical reliability.
Thesis Designed and built a real-time platform to rethink how university lectures work — a mobile-first PWA that gives professors live signal on what students actually understand, and gives students a voice in the room without having to raise their hand. Shipped end-to-end, from the WebSocket backbone to the classroom UX.
An end-to-end AI design platform for homeowners, agents, and architects. Upload a photo of any space, strip the clutter, apply a style, and get a client-ready render in seconds. Replaces expensive physical staging, eliminates rendering cycles, and cuts the time from idea to proposal.
An AI performance companion built for athletes who take their body seriously. Connects with Whoop, Garmin, and Strava to build a full picture of how training, recovery, and nutrition interact — then cuts through the noise with science-backed guidance on food, sleep, and readiness. Upload blood work or biomarker data and get clear, actionable direction. Built for people who want to perform at their best and actually understand why.