Sales one quarter, product the next, an outbound engine the quarter after. I'm a generalist by design — I go where the company is thinnest and make that part work.
Full-cycle: discovery, demos, trials, pricing, implementation. I've sat across from CXOs in four industries and closed. I also build the machine behind it — signal-led outbound across thousands of accounts.
Roadmap definition, MVP scoping, shipping. Clinical summary automation with OCR and LLM pipelines. Knowledge graphs from 180+ diagnostic parameters. I translate what customers say into what gets built.
An engineer's habit that never left. Queueing theory for EV battery swapping. Traffic assignment in C++. When a problem has structure, I find it before I start guessing.
An LLM workflow that generates account research, personalized messaging and meeting briefs. Wired into HubSpot and Salesforce.
Multimodal OCR plus LLM interpretation, turning messy clinical documents into structured summaries clinicians will actually read.
180+ diagnostic test parameters structured into JSON schemas, with a dashboard that reads a lab report and returns a summary.
Buying-trigger detection across 5,000 accounts and 25,000 prospects. 30% lift in replies, a fraction of the manual effort.
M/M/k queueing applied to peak-hour demand at swapping towers. Projected 20–30% cut in wait times.
Method of Successive Averages over a real road network, with Dijkstra routing and all-or-nothing assignment.
This is where the site grows every week. Short pieces on GTM, AI products, and the strange job of being the generalist in a startup.