I architect Generative AI systems that move from POC to production. I bridge stochastic AI research with deterministic engineering by building the Day 2 infrastructure enterprises need: governance, FinOps and cost modeling, identity-based security, and observability.
Education: M.Sc. in Artificial Intelligence, University of Jyväskylä (completed). Now pursuing a second M.Sc. in Data Engineering for AI at DSTI School of Engineering, Sophia Antipolis.
With 8 years in high-performance computing and industrial digital twins, I build decision engines, not chatbots. My philosophy is simple: AI is expensive and probabilistic. My job is to make it cost-effective and reliable under strict business rules.
- Enterprise RAG & Knowledge Management: Identity-aware retrieval with Row-Level Security, citation-first generation to eliminate hallucinations in finance and legal.
- Agentic Workflows: LangGraph state machines with deterministic routing. AI handles intent, code handles execution.
- Hybrid Intelligence: Anchoring LLM outputs to physics engines and business constraints, from my background in C++ simulation.
- AI & Orchestration: Python, LangChain, LangGraph, Semantic Kernel, LLM orchestration, RAG pipelines, Agent frameworks, Cloud AI platforms and APIs, Ollama, llama.cpp.
- Vector Infrastructure: Qdrant, Supabase (pgvector), Milvus, Pinecone.
- Databases: PostgreSQL, MongoDB, GraphDB - Neo4j, Redis.
- Backend & Systems: FastAPI (Python), Go (Golang), Rust, C++17, Apache Spark.
- Cloud & DevOps: Azure (Entra ID, Container Apps), Docker, Kubernetes, Terraform (IaC), GitHub Actions.
- Frontend: React, TypeScript, TailwindCSS, Streamlit.
- LinkedIn: linkedin.com/in/nibir-1
- Email: nahasat.nibir@gmail.com