Product · Security · Infrastructure

Selected work

Start with three featured projects or explore the systems I’ve built for agent safety, payments, observability, and more.

16 case studies1144 automated tests15 architecture decisions

01 / Product leadership

Funa Gigs

Hire nearby. Pay after approval.

A Ugandan marketplace for matching with local taskers, coordinating the job, and paying in UGX.

Read the Funa Gigs case study
Funa Gigs Android home screen with Instant Hire and service categories Funa Gigs Android worker-matching screen showing a nearby search in progress
Funa Gigs Android app · Official Google Play screenshots

02 / Defensive security

Cyber Dragon

Inspect suspicious messages without opening the link.

A browser workbench that explains the signs of mobile-money scams, lookalike domains, and USSD traps.

Read the Cyber Dragon case study
Cyber Dragon workbench with fraud inspection modes and built-in sample scenarios
Cyber Dragon · Live inspection workbench

03 / Developer tools

AgentFlight

Replay an AI agent’s failure in a test.

A recorder and replay tool for comparing AI agent runs without repeating live model or tool calls.

Read the AgentFlight case study
incident.flight.json

Recorded tool call · payment-api · 15m

{
  "p99_latency_ms": 1840.0,
  "anomaly": true
}

✓ Matching input replays the recorded response

What if the call asks for 30 minutes?
TrajectoryDivergenceError
window: expected "15m", got "30m"
AgentFlight replay · Simulated incident from the project README

Open source

Inside the systems

Each case study explains the problem, the design, and how to check the evidence.

  1. 01 AgentShieldCheck permissions before an agent acts. AI security 49 tests
  2. 02 FlowLedgerKeep balances exact through payment retries. Financial infrastructure 103 tests
  3. 03 StackPulseSee where a request spent its time. Cloud observability 83 tests
  4. 04 Auralis LocalSearch private documents on your own machine. Local AI 47 tests
  5. 05 ARGUS AIChallenge a proposed trade before it enters the simulation. Research tooling 77 tests
  6. 06 Nexus AegisAttach verifiable evidence to an agent’s proposed change. AI verification 87 tests
  7. 07 Cortex MCPKeep a history of what an agent knew and when. Agent memory 104 tests
  8. 08 PatchPilotTrace a failing test to a repair you can verify. Developer tools 92 tests
  9. 09 Synapse RouterEscalate to a stronger model when confidence drops. AI infrastructure 96 tests
  10. 10 SwarmWeaverCoordinate parallel work without competing file edits. Agent orchestration 70 tests
  11. 11 TruthGateCheck an agent’s claims against sources and code. AI verification 75 tests
  12. 12 AgriSignalDeliver farm advisories over short multilingual SMS. Offline-first systems 64 tests
  13. 13 Clinical Virtual Health PlatformCoordinate virtual appointments across web, mobile, and video. Digital health 5 tests