Platform engineer · three-time co-founder and CTO · Albany, New York
I have been building and running production systems since 1996, from 100 servers to 12,000, and from a first prototype to a signed acquisition. Four companies I helped build were acquired. Through Horizon Group I take on a small number of engagements where that experience is the point.
01 · Engagements
For funded startups between their first engineer and their first infrastructure hire. I own reliability, cost and the on-call story, and leave you with a system your next hire can run.
For health-tech teams that need patient records. I have taken an app through Epic registration, per-organization activation and signed-token production access, and I know where the documentation stops.
For teams that started on a hosted backend and outgrew it. I manage Postgres from standalone instances to RDS global clusters, obsess over row-level security, and build tooling that catches load issues before they become outages.
02 · Track record
| Years | Company | Role and what I ran | Outcome |
|---|---|---|---|
| 2023 to 2026 | Procare Solutions | Lead platform engineer. Moved the platform to multi-region Kubernetes, cutting deploys from an hour to five minutes and cost by half. Built multi-provider LLM routing with failover. | |
| 2023 | Readyset | Staff platform engineer. Designed and built the cloud platform that turned the product into a hosted service. | |
| 2020 to 2022 | Gemini Rockset divvyDOSE Chick-fil-A Supply | Senior platform and reliability roles on AWS and Kubernetes, several as a consultant. Cluster migrations and version upgrades, legacy infrastructure moved into Terraform, CI and deploy pipelines rebuilt, and on-call made sane. The common thread: enabling developers to ship with infrastructure they can rely on. | |
| 2018 to 2020 | Blockrize | Co-founder and CTO. Built the first crypto rewards credit card, from prototype to production infrastructure. Two patents granted. | Acquired by Gemini |
| 2017 to 2020 | MediaMath | Staff SRE and team lead. Cassandra, ScyllaDB, Redis and Postgres clusters across data centers and AWS. | |
| 2015 to 2017 | Agrilyst | Co-founder and CTO. Ran the engineering team and built the prototype that won TechCrunch Disrupt 2015. | Acquired by iUNU |
| 2012 to 2016 | Site reliability engineer. Integrated Admeld into DoubleClick, then led the rollout of Monarch monitoring from a quarter of Google's data center pods to all of them. | ||
| 2011 to 2012 | Admeld | Site reliability engineer. Rebuilt the infrastructure and automated new data center buildouts. | Acquired by Google |
| 2009 to 2010 | AppNexus | Manager of cloud operations. Teams in four cities, more than 1,000 servers. | |
| 2008 to 2009 | Yahoo | Manager of ad exchange operations. Grew the fleet from 5,000 to 12,000 servers and moved 7,500 of them to a new operating system with zero downtime. | |
| 2006 to 2008 | Right Media | Manager of Unix systems. Grew from 100 servers to more than 5,000. | Acquired by Yahoo |
03 · Building
A health record and assistant for people managing chronic illness, and the families who care for them.
hale.med ↗ Healthcare · infrastructureThe records layer under BentoBox: one connection to patient data across Epic and other systems.
pera.med ↗ Postgres · open sourceA self-hosted dashboard for reading and reviewing row-level security policies.
rlsee.io ↗ Kubernetes · open sourceShort-lived test environments from one YAML file, torn down on a timer.
mayfly.sh ↗ Networking · open sourceA per-hop network monitor that tells you where the outage is, not only that there is one.
pingway.net ↗ Databases · open sourceSchema versioning and migrations for Python teams, in the spirit of Flyway.
GitHub ↗I also make small things for the fun of it: cozy games at Littlest Bee, a one-page planner at narrow.day, and a ticket tracker called Twickets. The rest is on GitHub.