AI/ML Engineer - multi-agent LLM systems (LangGraph · Python) | Deep learning for time series

Currently building: an autonomous multi-agent coding platform on LangGraph: ~16-node workflow with specialized agent roles, Docker-sandboxed execution, semantic long-term memory (PostgreSQL/pgvector), provider-agnostic model routing, application-level AI safety, and verification-first design: 200+ unit tests, static-analysis gates, ground-truth benchmarks. Core modules are production-tested; public showcase coming with the proof of concept.

Earlier: an end-to-end RAG ingestion pipeline (PDF → chunking → embeddings → Pinecone).