Practice

Examples of what's been shipped — and the technology range behind each.

The work spans vertical SaaS, multi-agent systems, voice and text AI, AR and on-device ML, multi-tenant data, embedding-based matching, AI infrastructure, and production operability — across web, mobile, and TV.

Below are recent examples in depth. What you'd build with me almost certainly isn't on this list. The patterns are. Open any capability for the technology choices, the build approach, and the lessons.

Platforms

2 capabilities

Multi-Tenant Vertical SaaS

Schema-first, RLS in week one, GraphQL on top, multi-surface from day one.

Next.js 15React 19TypeScriptPostgresSupabaseRow-Level Security+4

Workroom & Production Displays

The shop floor's TV becomes the dashboard. Apple TV and Fire TV, live status, no input required.

tvOSSwiftUI for tvOSAndroid TVFire TVKotlinJetpack Compose+1

Mobile & AR

3 capabilities

3D Scan to Quote

From iPad in a customer's space to a structured, line-item quote in ninety seconds.

Swift 6SwiftUIARKitRoomPlanLiDARMetal+3

Voice & Text Operations

Run the business from anywhere. Same tools, same context, voice or text, picked by the situation.

Gemini Live (Bidi WebSocket)Tool-calling schemasDomain-grounded system promptsRefusal-on-uncertainty patternsSwiftUIiMessage-style chat UI+1

AR + On-Device ML for Mobile

Custom CoreML models, ARKit world tracking, HealthKit-grounded voice — all on-device, sub-100ms, no network required.

Swift 6SwiftUICoreMLVisionARKitMetal+6

Agent Systems

1 capability

Multi-Agent Orchestration

LangGraph topologies with self-assessment and verification gates between every handoff.

Python 3.12LangGraphFastAPIWebSocketPostgres state storeTyped inter-node contracts+2

AI Infrastructure

1 capability

Multi-Provider AI Routing

Claude for reasoning. GPT-4o for structure. Gemini for multimodal. Routed per task, with cost-aware fallback.

Claude (Opus, Sonnet, Haiku)GPT-4oGemini 2.5PerplexityPer-task routing tablesCost-aware fallback chains+1

Data

1 capability

Embedding-Based Matching with Re-Rank

pgvector candidate generation, structured re-ranking with hard constraints, per-result rationale.

PostgrespgvectorOpenAI embeddingsStructured filter layerOperator must-have rulesPer-result rationale

Need one of these for your product?

Tell me what you're trying to ship and what's getting in the way. If it's a fit, we'll scope a sprint.

Book a 30-minute call →