AI Engineer · Portfolio · 2026
>Multi-Agent BuilderAI Engineer · I build with Python, FastAPI, LangGraph, CrewAI, PgVector, Redis, and Azure OpenAI — shipping retrieval, orchestration, and eval pipelines that hold up in production.
I'm an AI engineer who founded and scaled eInfoChips' AI Studio from a solo proof-of-concept into a 9–11 member team — multi-agent LLM systems, RAG infrastructure, and adversarial LLM security testing across production platforms.
What I do
Most of my time goes into the unglamorous middle layer between models and product — retrieval, orchestration, evaluation, and the infra that keeps it all honest.
Designed supervisor-routing Team architectures on Agno, scaling a solo proof-of-concept into a 9–11 member initiative (NomAIzo™). Shipped CyberStudio with 6+ specialized GPT-4o agents, automating EU Cyber Resilience Act compliance gap assessments at ~70% accuracy.
Architected event-driven 14-microservice systems on FastAPI and RabbitMQ. Engineered high-throughput RAG pipelines with Docling OCR and dual, physically isolated PgVector knowledge bases utilizing hybrid semantic + keyword search and content-hash deduplication.
Established Langfuse and Garak observability and adversarial testing (90% security accuracy across 28 attack probes). Integrated Wiz for continuous code/container scanning and hardened APIs via Fernet/bcrypt encryption and Redis JWT blacklisting.
Selected work