Josip V.

ABOUT ME MD + PhD (medical ML) and a recent MSc in AI. I work hands-on as an ML/AI engineer, mostly on healthcare AI and LLM-native apps (RAG + agents + tool/function calling), and I’m comfortable taking things from idea → prototype → production. In the last couple of years I’ve been building and shipping: - LLM-native SaaS for customer support (text + voice agents): RAG with LlamaIndex/LangChain, dynamic function calling, and parallel tool execution for faster inference, plus backend-heavy full-stack work. - Cloud deployment & maintenance on AWS + GCP (Lambda/Cloud Functions, API Gateway, Postgres, etc.). - Voice-to-voice agents using LiveKit/Twillio and LLM APIs (OpenAI / Gemini). - Research and applied projects around clinical decision support, error checking, and safety with modern LLMs/LRMs. Recently, I’ve also worked on research projects using supervised fine-tuning (SFT) and RLVR with medgemma27n and GPT-OSS, focused on medical error checking and improving patient outcomes. Working freelance on: AI in healthcare / medical AI (classical ML + deep learning) RAG applications / RAG chatbots (LlamaIndex, LangChain, vector DBs, evaluation) AI agents (LangChain, LlamaIndex, Autogen) + tool/function calling Fine-tuning (SFT, preference/RLVR-style training), dataset curation, eval & guardrails AWS Lambda + API Gateway / GCP Cloud Functions deployment End-to-end ML pipelines (data → model → monitoring) Data analysis, statistics, biostatistics, visualization Academic & medical writing, literature reviews, consulting If you message me, the fastest way to start is: what’s the goal, what data you have (and where it lives), and how the output will be used (internal tool, customer-facing app, research paper, etc.). I’m happy to jump in either as “build the thing” or “fix/upgrade what’s already there.”