Hi there,
I'm Jeff.

1. 01

   Multi-surface products that feel alive

   Real-time, mapping, dashboards, and performance-driven UI across mobile, web, and native.

2. 02

   Agentic infrastructure that holds up in production

   Orchestration, evaluation, memory, and tool coordination built to keep multi-agent work reliable rather than chaotic.

3. 03

   Privacy-first architecture

   Minimal data collection, explicit handling, and sane defaults baked into the system from day one.

1. 01

   Instrument reality

   Observability and feedback loops so "working" stays working. Six ops dashboards, adversarial test suites, and live automation telemetry, not dashboards that only look good in a demo.

   See the PrayerMap case study
2. 02

   Respect users

   Privacy-first decisions, clear consent, and intentional data boundaries. On-device Metal shaders and CoreML pipelines where the data never has to leave the phone.

   See the projects index
3. 03

   Ship the boring parts

   Tooling, testing, migrations, and operational workflows that keep products stable. 46 edge functions, 198 migrations, and 183 Neurotoken tests are the un-glamorous reason shipping stays fast.

   Read the Neurotoken write-up
4. 04

   Optimize for clarity

   Simple interfaces, clean abstractions, and deliberate constraints. Three similar lines beat a premature abstraction every time.

   See the Claude Code Toolkit
5. 05

   Own the AI-assisted workflow

   I develop with Claude Code as my primary environment. The AI handles implementation velocity; I own architecture, verification, and quality. Every Co-Authored-By trailer in my commit history is a transparency practice, not a crutch. When the agent gets it wrong, I catch it, reject it, and redirect.

   See the GitHub commit history

1. Institutional finance

   Years inside a regulated, risk-first environment. Trained to make decisions under uncertainty and quantify what can go wrong before it does.

2. Founder and operator

   Built and ran businesses with real customers, real payroll, and real constraints. Learned that discipline beats novelty and that systems compound.

3. Shipping engineer

   Full-stack engineer and AI workflow architect shipping production systems across mobile, web, native, and orchestration infrastructure.

Open to roles and collaborations where the work is high-signal: product engineering, AI workflow infrastructure, and teams who care about craft. Performance, reliability, and humane UX are non-negotiable, and I am most useful where ambiguous ideas need to become shipped systems under real constraints.