I build production-grade GenAI systems — not demos.
RAG, agents, and ML pipelines with grounding, evaluation, and deployment in mind.
What I build
Three lanes. No role-shopping. Inspectable systems.
GenAI Systems
RAG pipelines, grounding + citations, agent orchestration, API-first design.
ML & Data Pipelines
Ingestion → ETL → indexing/training, batch workflows, cloud deployments.
Evaluation & Reliability
Offline eval, regression tests, latency awareness, fallbacks & failure modes.
Selected work
A few systems that show how I think, build, and ship.
GENAI • RAG
ROHbot — Portfolio RAG Assistant
LLM assistant that answers questions about my work using retrieval + grounded responses with sources.
CV • DIFFUSION
Sketch Studio — Sketch → Photorealistic Portraits
Stable Diffusion + ControlNet + BLIP pipeline to preserve structure and improve prompt adherence.
Technical writeups
Deep dives after you’ve seen the code.
Let’s build something real
I’m interested in roles where I can design and ship real ML and GenAI systems.