I work on Microsoft's Forward Deployment Engineering (FDE) team, partnering with the world's most advanced organizations on high-stakes, high-complexity AI deployments where edge cases are the norm. My focus: building things that matter at scale, always grounded in a strong ROI-driven mindset.
- Real-Time Voice AI — building HIPAA-grade speech pipelines that actually hit sub-second latency
- Agentic Protocols — MCP, A2A, and figuring out how agents should talk to each other
- Memory for Agents — making context persist across sessions without losing its mind
- LLMOps — eval harnesses that go beyond vibes, tracking latency, cost, alignment, and quality
- AI × Healthcare — where all of the above gets really interesting (and really hard)
|
|
|
I write about the stuff I build and the problems I run into along the way.
- The Hard Part of Software Was Never Coding — the craft moved upward, from typing code to steering systems
- AI-Powered Clinical Knowledge Stores — turning messy clinical docs into searchable knowledge
- Prior Auth + GenAI — how we cut turnaround by 70% with agentic AI
- GenAI Production Challenges — real blockers, not theoretical ones
- LLM/SLM Evaluation — building eval that you can actually trust
- Document Management + Azure AI — OCR, NER, and making unstructured data useful
- Built AI HLS Ignited at Microsoft — the HLS GenAI accelerator used across providers, payors, and life science orgs
- Teach cloud computing & AI at Northwestern (MSAI program)
- Led MLOps at Concentrix — real-time + batch ML for Fortune 500s
- Architected the MLOps accelerator at Levi's — Python SDKs, computer vision, product classification
Want to build something together?
Always down for open-source collabs, weird side projects, and working with people who ship.
