I will fix, debug or improve your llm or chatgpt API integration

F
filipj93
F
filipj93
Filip Jukic

About this gig

Your LLM feature almost works? I'll find out why and fix it.


Typical problems I fix:

  • API calls that fail, time out or return broken JSON
  • tool calls and structured output the model ignores
  • RAG chatbots that pick the wrong documents or make things up
  • prompts that work in testing but not with real users
  • high token costs, slow responses, streaming bugs

Send me the code and an example of what goes wrong. I reproduce it, fix it and explain what changed. Works with OpenAI and OpenAI-compatible APIs; Python preferred, TypeScript and Java fine.


About me: I build LLM systems that run in production: AI coding of medical documents for hospitals, and a report generator that runs about 170 LLM calls per report with retries and caching.

Get to know Filip Jukic

Filip Jukic

AI Engineer for LLM agents, private LLMs and Python backends

  • FromAustria
  • Member sinceSep 2022
  • Languages

    German, Croatian, English, French
I build LLM systems that run in production: agents with tool calling, RAG over your own documents, and self-hosted open-weight models (llama.cpp, sglang) so sensitive data stays on your own servers. Right now I'm building AI coding of medical documents for Austrian hospitals and an AI report generator for energy audits. Background: MSc in financial and actuarial mathematics (TU Wien), about 9 years as a software engineer at IBM and Zurich Insurance, in insurance and finance. Python, Rust, Java, TypeScript. Clear scope, fixed price, clean handover.