About

I am an AI Forward Deployed Engineer at IBM. The work starts where most deployments end: once the platform is live, I work directly with the client to build the custom solution they actually need on top of it, then move that solution onto their own systems so it runs cleanly in their environment rather than only in ours.

That last part is the whole job. A solution that only runs in a demo environment is not finished, so most of the work is learning a client’s constraints well enough to design against them from the start — and then making the handoff boring.

Underneath it is the platform work: RAG pipelines on IBM Granite and Llama serving enterprise IT and HR support agents, inference tuning to keep latency and cost in budget, and the evaluation harnesses that decide which model actually ships.

I came to it from both sides of the stack. I finished a dual B.S. in Computer Science and Electrical Engineering at Penn State in 2026, which is a long way of saying I am as comfortable debugging an RTOS scheduling conflict across three STM32s as I am profiling P95 latency on a model endpoint.

Outside of work I am still in the lab and the garage: electrochemical biosensor automation at Penn State’s E-Bio-E Laboratory, and an air hockey robot that has to see, predict, and move in under 20 ms. The through line is correctness under real constraints — whether that constraint is a P95 budget on an inference endpoint, a client’s production environment, or a stepper motor that will not wait for a slow frame.