Everything we do is only as good as our data.
I'm Jon. I lead AI, machine learning, and engineering teams that ship — across healthcare, life sciences, legal services, transportation, and financial services.
I started in data architecture, earned a master's in predictive analytics at Northwestern, grew a global engineering org from 10 to 30+ across three countries, and now lead Applied AI Delivery at DLA Piper. I've stayed hands-on the whole way — and in this field that's not nostalgia, it's the job. You can't set direction on AI systems you've never built, and you can't govern what you don't understand.
Twenty years working in the data — not above it — has made me stubborn about one thing. AI can help you clean data, profile it, reconcile it. What it cannot do is work around it. Feed a model bad data and you don't get a broken answer you'd notice; you get a confident one, faster and at greater scale than you could produce it before.
So you still have to know what data quality actually means — completeness, lineage, timeliness — before you serve any of it to a model. That was always true. AI just raised the cost of getting it wrong.
Which is why I build the unglamorous parts first: the pipelines, the MLOps, the evaluation harnesses, the governance. Then the impressive parts actually hold.
Outside the day job, I run AI Force Consulting, helping small and mid-sized businesses build responsible AI strategies, and I ship consumer mobile products like Scavenger Go to keep my 0→1 instincts sharp.
Responsible AI strategy for small and mid-sized businesses. Most SMBs don't need a research lab — they need a clear-eyed plan for where AI actually pays off, what data it requires, and how to deploy it without creating risk they can't govern. That's the work.
A location-based mobile app taken from idea to working product in weeks, not quarters — every storefront is a clue, with live leaderboards and real prizes.
This is the work I want more of: helping small teams — and companies with no IT department at all — put something real in front of users fast. You don't need an engineering org to ship software people actually use.
Azure OpenAI applications for document drafting, regulatory analysis, and structured intelligence at one of the world's largest law firms — built on LangChain, LangGraph, Azure AI Search, and LangFuse.
A greenfield Snowflake, Azure, and Power BI platform powering AI-enabled analytics and automation for a multifamily real estate investment firm overseeing ~$1.6B in AUM — cutting annual operating costs by more than $250K.
Think Yelp for logistics: ratings and real-world operating intelligence on 3M+ commercial locations, so carriers knew what they were driving into before they got there. I directed the ML development and cloud platform engineering — plus the MLOps foundation that got those models into production.
Modernized a transportation SaaS platform that was bleeding customers: eliminated recurring outages, moved releases from twice a year to monthly, reversed attrition, and helped double revenue.
Semantic search, built with agentic intelligence — agents that plan the query, chase down sources, and reason over what they find. Enterprise-grade retrieval, the kind that normally carries an enterprise-grade invoice, at zero cost to the people using it.
Everyone should have real AI search at their fingertips. It shouldn't sit behind a paywall for the few who can expense it.
Hiring, partnering, or trying to figure out where AI actually fits in your business? My inbox is open.