Before this, I spent seven years as an ML engineer in healthcare with a focus on automation. At a medical-coding AI company, I built the PySpark pipelines on Databricks that moved 10M+ cases into production and the SageMaker services that predict ICD/CPT codes at claims scale. Before that, at a radiology AI startup, I built the Faster-RCNN models that detect breast cancer in mammograms — 87% AUC on a thousand-mammogram test set, deployed directly into PACS workstations.
Today I'm an ML engineer at Cellhub, working at the intersection of AI and mobile networks. On the side, I help the small and mid-size businesses most AI vendors ignore figure out which AI is actually worth buying. Most of those problems trace back to the same gap: the people building AI and the people who have to sell, buy, or operate it rarely speak the same language. That translation is the job.
How I got here
The last few years, I've had a front-row seat to AI from the inside. As an ML engineer, I watched my own field change under my feet — software teams shifting from writing code by hand to running orchestras of LLM agents that write it for them. From where I sat, knee-deep in production ML in an industry where "it doesn't work" is never a minor inconvenience, the shift was real. Not a demo.
The lightbulb came from a different room entirely. I was helping my uncle with his construction company — voicemails, paper invoices, the owner's head as the only source of truth. The same tools I'd spent years using to automate medical coding and radiology workflows could change how a business like his actually runs. Not in a five-year roadmap. Now.
That's where this work sits. I write about the AI shift unfolding across telecom and mobility — the industries I'm inside now, watching the same automation patterns I shipped in healthcare show up a few years behind. And I help the small and mid-size businesses most vendors ignore cut through AI theater — owners who don't need a lecture on transformers, they need someone who can tell them what's worth a check and what's not. Seven years of shipping production ML earns me the room to do both.
What I write about
- AI across telecom and mobility — what's actually getting deployed inside carriers and across the physical networks we move through, which deals close, and where the regulated edges live
- AI sales architecture — the motions that work when there's an operator in the room who can tell the difference
- Small business AI — practical patterns that survive a real P&L, not theoretical frameworks
- Healthcare AI, from inside — receipts from seven years of shipping medical-coding and imaging ML into production where failure isn't a KPI miss, it's a missed diagnosis
Who this is for
Readers are small and mid-size business owners trying to figure out which AI actually pays for itself, operators and product people tracking how AI is reshaping telecom and mobility, and the GTM leaders selling AI into regulated industries who want analysis from someone who's been on the other side of the buy. If that's you, I write for you.
Writing series
- Carrier Watch — What's really happening inside the networks and mobility stacks everyone talks about.
- Field Notes — Observations from conversations with operators and SMBs navigating AI decisions.
Off the clock
Based in Boston. When I'm not at the keyboard, I'm usually driving through the dunes of Cape Cod — still the most reliable reset I've found. Currently reading Dungeon Crawler Carl, which turns out to be a masterclass in how pure voice can carry a ridiculous premise.
Get in touch
The fastest way to reach me is on LinkedIn. If you're working at the intersection of AI and mobility, selling AI into a regulated industry, or running a small or mid-sized business trying to separate the AI that pays for itself from the AI that doesn't — send me a note.