Backend and AI engineer with 5+ years building production systems across fintech, blockchain and B2B SaaS. I co-founded Borebase and built it end to end - from architecture and domain logic to billing, deployment and onboarding. A backend engineer who ships the frontend too: async Python and PostgreSQL at the core, React/Next when the product needs it, plus multi-chain integrations and multi-agent RAG. I've shipped systems for products reaching 100K+ users.
I build production backendsAI-powered systemsfintech platformsblockchain infrastructure
Originally from Latvia, Riga, with 5 years of study and work in Netherlands, Eindhoven. I've been paid to build software since 2021, and I've been teaching it to kids on weekends for almost as long. 500 of them by now, ages 7 to 18.
Teaching was never just a side gig. It's where I learned that if I can't explain it to a 9-year-old, I don't understand it yet. That habit is most of my engineering.
I'm the one who reads the standard before writing the module, digs until the whole system makes sense and not just the part I need, and stays with the problem after the interesting part is over. Payments, agent systems, on-chain rails, geotechnics. The domains keep changing. The method doesn't.
If there's a problem nobody wants to own, I'm usually the right call.
Any player opens a lobby and the bot holds the queue until 10 join, then splits the sides with a custom Elo model, deliberately varying the pairings so the same ten never get identical teams, and moves everyone into team voice channels. With no game API to read results from, players call the winner themselves: the team with more reactions takes the match, and every rating is recomputed and persisted in MongoDB.
An AI CLI for data engineers & database architects. Turns natural-language requests into SQL, so you can query and explore databases without hand-writing queries.
CodeMonitors 12+ exchanges (REST + live WebSocket order books) and DEX aggregators across chains, detects cross-venue spreads and validates each for real executability - depth, network routes, bridge feasibility (deBridge), token safety (GoPlus) - before tiered Telegram alerts.
CodeProbes router / ISP / DNS / VPN in parallel and uses consensus and hysteresis to infer the most likely failing layer, packaged as a single static binary with a live terminal dashboard.
A dual-bot system (admin + storefront) with full catalog/order CRUD, statistics, broadcasts, discounts and multi-wallet management. Detects bot bans and fails over to a backup token instantly, with on-chain USDT payment verification.