Hello, I'm Priyanshu.
I'm a AI/ML Engineer with 5 years of experience building and deploying models.
W. Edwards Deming once said “In God we trust; all others must bring data.”
I build AI aligned with human values—systems that work with people, not instead of them. In practice, that means human-in-the-loop design, clear acceptance criteria, and evaluation pipelines where performance, safety, and drift stay visible and improvable. That's where I come in: turning complex information into the “aha!” moments that actually matter—so from retrieval and fine-tuning to deployment and monitoring, messy knowledge becomes dependable answers and better decisions.
Recently at BlackRock and AIG, I've been scaling retrieval, evaluation, and automation so the models stay grounded and useful over time.
I like owning the whole flow: data pipelines → retrieval → model choice/fine-tuning → deployment → monitoring. I care about aligning AI with human values in practice—transparent reasoning, trustworthy guardrails, and failure modes you can actually see and fix. I prefer strategic simplicity: small, composable choices that help teams move fast without breaking trust.
Tech I use : Python, SQL, PyTorch, Hugging Face, LangChain, RAG pipelines, vector DBs, Databricks, AWS (and FastAPI/Docker for services), evaluation harnesses and monitoring, plus MCP-style tool integrations when they help close the loop.
Outside work, I'm the kind of person who'll play poker (mostly to lose and learn), train from zero to a half-marathon, and keep showing up to salsa even though my moves are still… aspirational. Keep it human, keep it measurable, keep improving.