Featured case studies
Work most relevant to applied AI, platform, and founding-engineer roles
These are the clearest examples of turning ambiguity into shipped systems, production workflows, and business-useful software.
Multi-agent operations layer
My OpenClaw company system
Built an internal multi-agent operations control plane for coordinating real business work across agents, memory, scheduled workflows, messaging, browser automation, metrics, and guarded execution.
- Designed specialist agents, durable state, command controls, and scheduled operating loops
- Coordinated product, release, research, monetization, and operational workflows across properties
- Built applied-AI operations infrastructure around real business workflows, approvals, memory, tools, and execution loops
AI app launch cockpit
Built a customer-facing AI app planner, private-preview builder, and ownership cockpit for businesses that want real web apps they own.
- Turns plain-English app ideas into reviewed launch plans and scoped private preview requests
- Coordinates customer-owned GitHub, Vercel, and Supabase setup with token, audit, and workspace boundaries
- Includes readiness checks, app health, domains, ownership maps, audit logs, and production approval gates
Applied AI infrastructure
Built a distributed local-LLM compute system that lets volunteers process open biomedical literature into citation-backed evidence candidates for researcher review.
- Designed signed work-packet flow, local worker execution, validation, consensus, provenance, and public impact surfaces
- Built Windows desktop worker controls for local model readiness, pause/resume, idle/battery/CPU limits, and activity visibility
- Added trust/safety layers across enrollment, download verification, moderation, admin controls, abuse monitoring, and public metrics
Trusted AI knowledge infrastructure
Built Forge to turn approved business documents into cited, reviewable knowledge that teams can trust before it feeds AI, search, or customer-facing answers.
- Designed document/source ingestion, evidence-backed extraction, review states, confidence signals, exports, and knowledge-credit foundations
- Focused the product around citations, source provenance, review workflow, and operator trust instead of generic document upload
- Built the trust layer needed before business knowledge can safely power AI, search, and customer-facing answers
Production AI answer delivery
Built Bridge as the delivery layer for approved knowledge: cited answer APIs, widgets, chat, releases, analytics, evals, and production trust controls.
- Built knowledge bases, retrieval, release lifecycle, API keys, widgets, answer/chat/retrieve APIs, analytics, evals, and RBAC
- Validated hosted launch readiness with migrations, indexing, release rollback, widget origin checks, API retrieval, and RBAC negative tests
- Delivered RAG infrastructure as a usable product with citations, controls, integration surfaces, and launch-readiness validation
Service-business sales workflow
Built DealOps to help service businesses move quoted work from proposal to approval, payment, and kickoff without losing the commercial thread.
- Built buyer/operator workflow across proposals, follow-up, deposits or full payment, and kickoff-ready handoff
- Strengthened payment reliability with Stripe webhook correctness, idempotency, and audit scripts
- Turned messy service-sales operations into a clear product workflow with buyer and operator surfaces
Controlled customer follow-through
Built Tavrio to recover missed leads, keep proposals moving, and improve payment follow-through with configurable review and sending controls.
- Designed follow-through workflows around missed leads, proposal follow-up, payment reminders, and operator review posture
- Balanced automation with visibility and control so teams can start narrow before expanding
- Balanced revenue recovery, automation, and operator control in a customer-facing workflow product
Founder / principal engineer
Built and led a patented market-research and engagement platform from concept to scale, raised $500K+, and supported a 20k+ participant user base.
- Founder-level ownership across product, engineering, operations, and growth
- Two issued U.S. patents in predictive modeling systems
- Adds founder-level operating history, long-term product ownership, and scaled user growth
Regulated product engineering
Tour Trader Pro / Player Options
Built regulated wagering products across Flutter, React, PHP, SQL Server, and Azure, including two GLI-33-certified products.
- Took Tour Trader Pro from idea to MVP in under two months
- Owned real-time pricing, integrations, CI/CD, monitoring, and ops workflows
- Shipped under compliance, reliability, and audit pressure in a regulated product environment
