Build AI that
survives production
Engineering judgment for production AI.
I take AI from the business problem to a system people use every day: strategy, product, engineering, delivery and adoption. One person accountable for the whole chain.
Stride | Plaid | MX | Barclays | UDR | Perlman Medical Group

Real outcomes.
MBAAWS Solutions ArchitectAWS AI PractitionerPSPOPSM II
How I work
The whole AI value chain, end to end
Most AI projects stall in the handoffs: strategy to product, product to engineering, engineering to delivery, delivery to the people who have to use it. I own all five stages, so nothing gets lost between the people who decide, the people who build and the people who use it.
Strategy
Discovery with executives and the people doing the work. Where the money is, what the ROI case looks like and what to build first. Sometimes the answer is not AI, and I say so.
Product
Requirements a team can build against, with the acceptance bar set up front: golden datasets, evaluation thresholds and where a human reviews.
Engineering
AWS serverless architecture, agents, RAG and model choice sized to the job, and the code itself. Governance a security team can sign.
Delivery
The long last mile: security reviews, SSO, load testing, cost limits and launch. I stay until it runs in production.
Adoption
Training, playbooks and measurement, so people actually use the system and the numbers show it paid off.
Recent work
Shipped, not demoed
AI content review at Stride
Product lead for CAIRT, an AI platform that reviews curriculum for grammar, bias and WCAG accessibility, integrated into Stride's AI course-delivery platform.
- 4 days to 2 hours QA time (on track)
- 90% detection, 5% false positives (evaluation)
- $0.37 per lesson
An AI SDR in production
Built by a self-contained pod reporting to the CEO of a marketing analytics agency, on a durable agent runtime with human approval and cost limits.
- 3,000+ prospect conversations
- New AI line of service launched
Is this medical group ready for AI?
Led a five-person AI pod inside HIPAA, EHR and FHIR constraints. Shipped a referral tool and an honest readiness assessment that leadership put on its roadmap.
- Built to cut referral time 60%
- 30 referral coordinators
One view from ship to depot
A branded, AI-first shipment platform for a carrier: multi-tenant AWS serverless, an AI morning briefing, and human approval on every action.
- Ocean, rail and truck in one journey
- 400+ commits
Content series
Production AI, explained
11K+ subscribers on YouTube. How AI systems get scoped, built, evaluated and run in production, broken down for the people who have to make them work.
Subscribe- In the DayWhat building production AI looks like day to day.
- In the TrenchesProduction AI case studies.
- The Production ModelArchitecture and mental models.
- Live Architecture ReviewsReal decisions, reviewed live.
The playbook
The AI Production Framework
10 pillars and 140+ principles for taking AI from prototype to reliable, governed, cost-bounded production on AWS. It is the playbook behind every engagement, published so you can judge the thinking before you talk to me.
Read the frameworkBackground
15 years of building, end to end
Strategy · Product · Engineering · Delivery · Adoption
I grew Net Synergy to 50 people over 14 years by running the same play on every engagement. I sat with clients like Plaid, MX, Barclays, C.H. Robinson and UDR to work out where the money was. I turned that into a product my teams could build. Then we engineered it, shipped it and stayed until it ran: more than 100 projects, including financial data platforms automating over a million accounts.
Systems only pay off when people can run them, so I built LT3 Academy to train those people. It took adults without college degrees into tech careers through Agile, SQL, Python and automation. It became the smallest company in the country to earn U.S. Department of Labor apprenticeship approval, with a 90% certification rate and 72% job placement.
Today I run that same chain for AI. Strategy, product, engineering, delivery and adoption, for organizations that need AI to hold up in production every day.
Have an AI initiative that needs to reach production?
Tell me what you are trying to ship. I will tell you honestly what it takes, and whether AI is the answer.
Start a conversation