Maksim Krauchyk
Full-stack developer focused on LLM-powered products. I build end to end: data model, API, OpenAI/Anthropic logic, React front end, deployment and documentation.
About
Full-stack developer focused on LLM-powered products. I build end to end: data model, API, OpenAI/Anthropic logic, React front end, deployment and documentation.
- Two client web platforms I delivered are live with 1,180+ and 1,030+ active users, including online payments, hosting and transactional e-mail.
- I lead Switchyard, an open-source LLM gateway in front of 25+ model providers (about 400 of 578 commits, 470+ automated tests).
- LLM integrations and RAG pipelines for clients; automation with Make, n8n and Zapier.
Stack: TypeScript, React, Node.js, Python/FastAPI, PostgreSQL, Docker. Based in Gdansk, Poland (EU time zone), English B2+, Polish C1.
- 3 years' experience
- Joined October 2026
Skills· 15
Languages
- EnglishAdvanced
- PolishFluent
- RussianFluent
- BelarusianFluent
Industry experience
- Software & Internet
- Computers & Electronics
- Financial Services
Work experience
Created and maintained an OpenAI- and Anthropic-compatible API in front of 25+ model providers. Implemented routing by measured latency and success rate with automatic fallback and “decide mode” for selecting the best model per request; added tool calling repair for broken JSON. Reduced agent context size per benchmark task from ~69k to ~7k tokens while keeping results; added a CI setup with 470+ automated tests and supported contributions from external developers.
Delivered two production web platforms from requirements to deployment and support; both are live with 1,180+ and 1,030+ active users. Built LLM integrations with OpenAI and Anthropic APIs and implemented RAG pipelines using vector search for client use cases. Integrated online payments (bepaid) including callbacks and transaction state handling. Set up hosting, DNS, SSL, and transactional email (SPF, DKIM). Worked with non-technical clients in English and Polish, translating business needs into scoped tasks and providing progress reports.
Built an asynchronous Python platform that processes messages from communities totaling 300,000+ members. Operated in production with structured logs and metrics, using PostgreSQL as the primary datastore.
Education
Hire Maksim Krauchyk today
Post a job and invite Maksim, or contact them directly. No hidden commission.




