Data & AI-Application Engineer
Production AI Systems, Built to Ship
LangGraph agents · RAG · Data pipelines
During a Master of Data Analytics in Canada (GPA 4.13/4.3), I designed, built, and tested projects end to end: a ~20-node LangGraph job-search agent, a LangChain RAG pipeline, a Go/MQTT IIoT backend, and a geospatial climate analysis of 275,156 trees. PGWP-eligible, no employer sponsorship required.
Recent Hands-On Work
From data to model to shipped — end to end.
LLM Agents & Orchestration
LangGraph evaluation pipeline
A ~20-node LangGraph StateGraph that takes a job description and produces a fit score plus a tailored resume and cover letter; LLM calls run across the Claude and OpenAI APIs with structured outputs and graceful fallback, covered by 460+ pytest tests.
RAG & Retrieval
Retrieval-augmented generation with source attribution
A LangChain RAG pipeline: document chunking → vector retrieval (FAISS/Chroma) → LLM synthesis, with answers attributed back to the source passages so they can be verified.
Data & IIoT Backends
Edge ingestion to time-series storage
The Matrix-Sync IIoT backend: a Go edge agent publishes over MQTT/Mosquitto, a TypeScript/Node ingestion service writes to PostgreSQL/TimescaleDB, and the whole stack runs via Docker Compose.