Track Record

Outcomes over outputs.

Five client engagements. We ran engineering on every one.
All of it on the public record.

Cyndra AI cyndra.ai CTO · 2026–present

Autonomous AI employees on hardware the customer owns. Live in hours, theirs the moment they boot.

Context

Businesses wanted AI team members. Nobody was actually selling one. Cyndra provisions autonomous AI employees that run on hardware the customer owns: their machine, their data, no landlord.

What we did

Yudame holds the CTO seat. We built the provisioning stack: configurable personas with persistent memory and a local skill layer, all of it working through the same email and chat tools the human staff already use.

Result

Each one is working within hours and answers to a human supervisor, like any other hire. The customer owns the machine it runs on, and everything it learns stays on it.

  • AI agents
  • Model Context Protocol
  • On-premise deployment
PsyOptimal psyoptimal.com CTO · 2025–present

Twenty-five years of US Military assessment methodology, online for the first time. Now it's in front of thousands of athletes nationwide.

Context

The Oklahoma City Thunder's team psychologist had spent 25 years building assessment methods for the US Military. All of it lived on paper.

What we did

We built the whole thing: architecture, backend, data layer, deployment. At the center is a real-time scoring engine that ranks a psychological profile against team, league, and sport-wide norms as the answers come in.

Result

It runs nationwide in English and Spanish, for thousands of students, athletes, and businesses. The Thunder won the 2025 NBA Championship in the platform's first full season.

  • Python
  • Django
  • PostgreSQL
  • Real-time scoring
Fuse fuse.is CTO · 2024–2026

From 0 to 1 in six months. The multi-tenet marketing intelligence platform now runs for 400+ growth teams.

Context

Marketing teams were copying numbers out of Google Analytics, Google Ads, and Meta into spreadsheets by hand. The strategic questions still went unanswered.

What we did

We architected it on PydanticAI with MCP-pattern data orchestration months before the industry settled on the approach, and built the verification layer that keeps every AI answer tied to real data.

Result

We hired and trained the dev team to keep scaling after we left. Today 400+ growth marketing teams run on it, Ziff Davis among them. The best work we do stops needing us.

  • Python
  • Django
  • PydanticAI
  • Anthropic Claude
  • Model Context Protocol
  • Data pipelines
  • Next.js
  • Cloud scaling
Official → Bumble Inc bumble.com Head of Engineering · 2022–2023

We carried a startup through its acquisition by Bumble Inc, rebuilt the engineering team mid-flight, and shipped every sprint doing it.

Context

Official, a relationship app serving 2M users (100k DAU), was headed into acquisition by Bumble Inc (NASDAQ: BMBL). The founder needed steady technology leadership through the exit.

What we did

Led all technology through the acquisition, then rebuilt the entire 8-person engineering team in six months and integrated data pipelines that cut report latency by 75%.

Result

Every sprint commitment met through the transition. The React Native app was handed to Bumble's engineering organization, aligned with the listed parent's operational and compliance standards.

  • React Native
  • Data pipelines
  • Cloud scaling
  • Team building
Simplenight simplenight.com Cache Architect · 2021–2024

Pricing for 700,000+ hotels worldwide, kept fresh and served in under 100 milliseconds.

Context

A global travel distribution platform whose hotel rates change constantly, across 191 countries. The existing system couldn't keep pricing fresh at that scale.

What we did

We designed a Redis caching layer between the GDS feeds and the consumer-facing API. It chews through 70GB of rates and availability a minute in real time, geospatially indexed across the entire inventory.

Result

API response times dropped 85% to under 100ms, keeping displayed prices accurate and booking flows fast enough to convert.

  • Python
  • Django
  • Redis
  • ETL systems
  • Geospatial indexing
  • Cloud scaling
Contact

Start a conversation

Reach Tom Counsell on LinkedIn.

Connect on LinkedIn