CTO @ Delivr.ai | Deterministic identity resolution & person-level intent
I'm the CTO of Delivr.ai. We resolve B2B intent down to the person: who is researching what, deterministically, keyed on hashed emails. Most of my time still goes into the Rust and Go infrastructure underneath.
- Rust data services (DataFusion, DuckDB) that took over from Snowflake, Spark, and Airflow. Billions of signals a day.
- An identity graph holding hundreds of millions of person profiles
- The intent API: 1.2B+ person-topic pairs, reimported into DynamoDB daily behind an atomic pointer swap
- An AI-assisted dev loop where agents write the first pass and humans own the merge
- Director of Engineering at Mediafly: strategic products contributing sustained $4M ARR, built the company's first nearshore engineering team
- Principal Engineer at InsightSquared: ETL platform behind 500% customer growth
- Performance Engineer at Intel: open-source benchmarking tools, 2x Division Recognition Award recipient
Languages
Data & Infrastructure
Web
-
Agents implement, humans gate on August 31, 2026
Agents write most of our first-pass code. Humans review every diff and perform every merge. The interesting part is not the agents. It is the harness that makes their output safe to ship.
<li>
<a href="https://www.delivr.ai/engineering/rebuilding-the-intent-graph-every-day"><b>We rebuild the intent graph from scratch every day</b></a> on August 25, 2026
<br /><i>Over 1.2 billion person-topic intent signals across roughly 27 million people and 2 million companies, thrown away and rebuilt whole every 24 hours. For a signal that decays in days, freshness is not a cadence choice. It is an accuracy claim.</i>
</li>
<li>
<a href="https://www.delivr.ai/engineering/view-through-is-a-join-not-a-model"><b>View-through is a join, not a model</b></a> on August 24, 2026
<br /><i>We put the same identity pixel inside the ad slot and on the advertiser site, then measured what display actually does. The typical payoff is a return visit three days later, which is exactly what click math cannot see.</i>
</li>
</ul>
Ouroborai - AI-powered B2B SaaS development platform that enables non-technical founders to deploy production-ready applications and receive ongoing AI-generated features.
EngineerDNA - Local-first, open-source engineering metrics platform. An alternative to SaaS tools like LinearB and Jellyfish that runs locally, targeting engineering managers of 5-50 person teams.
An open-source benchmarking suite for x86_64 Linux systems. SPET automates performance testing across compute, memory, HPC, and web workloads for system administrators and performance engineers evaluating server hardware.
Tech: Python, Bash, C++
Production-ready fullstack application template with authentication, database configuration, API structure, and deployment setup.
Tech: React, Node.js, Express
Opinionated Node.js project template with TypeScript, testing, linting, and CI/CD configurations.
Tech: Node.js, TypeScript
Click to expand
Always up for talking identity graphs, intent data, or DataFusion war stories.
This profile automatically updates every 3 hours | Last refresh: Saturday, September 12 at 1:01 PM CDT