Embedded discovery
Technical work shaped beside enterprise engineering teams, against real environments, constraints, and adoption behavior.
← Recruiter fast track
Forward deployed engineering / AI products / customer delivery
I turn ambiguous user problems into shipped systems.
A decade working directly with enterprise users now combines with zero-to-one product engineering: discover the real workflow, build across the stack, observe actual behavior, and convert recurring patterns into platform improvements.
10+
years customer-facing
140K+
TypeScript lines shipped
iOS
live consumer product
60%
enterprise adoption reached
01 — Why the fit is real
Evidence across employment and independent builds.
Full-stack delivery
React Native, Next.js, TypeScript, Go, Postgres, APIs, AI, commerce, deployment, and operations owned end to end.
Product feedback loops
Recurring user behavior becomes tooling, architecture, product direction, and measurable adoption—not a one-off workaround.
Agentic systems
MCP, grounded retrieval, durable memory, tool-driven agents, inference, evaluation, and production telemetry.
02 — Direct proof
Four paths. No scavenger hunt.
Product
Bricktelligence
An AI-native companion shipped from problem definition through mobile, backend, AI, commerce, and operations.
Inspect →
Pattern
The Archivist Stack
A repeated product architecture extracted, secured, tested, documented, and open-sourced.
Inspect →
Judgment
Decision Ledger
Consequential choices documented with rejected alternatives, costs, and evidence.
Inspect →
Delivery
Professional Experience
Enterprise platform adoption, production observability, technical discovery, and customer outcomes.
Inspect →
03 — Interview walkthrough
Show the reasoning, not a feature tour.
Ask me to walk from a user problem to a shipped AI-native product. I can show the product decision, interface, data model, event flow, agent tools, deployment, and the architecture I extracted after building the pattern three times.
01
Begin with a real user decision
A collector does not need another catalog; they need help deciding what to build or buy.
02
Model the behavior, not the pitch
Events, lifecycle state, durable facts, and aggregates capture what actually happens over time.
03
Give the agent grounded tools
The assistant queries collection history, pace, progress, and structured set data instead of improvising.
04
Turn learning into product architecture
The recurring entity/event/memory substrate became an inspectable open-source starting point.
The short version
I do not stop at understanding the system. I ship the capability and stay close enough to learn whether it worked.
Download matching résumé ↓