Workflow first
The work starts with the broken operational path: inputs, handoffs, edge cases, review points, and the output people need.
AI workflow implementation proof
Shipped systems you can inspect before a sprint.
Kenneth builds product surfaces, backend workflows, agent loops, deployment paths, and verification gates for operational AI systems.
The work below is organized by capability, not vertical: agent operations, CRM automation, public AI interaction, and customer-facing workflow design.
Selected work
Capability proof
Agent operations system for shipping workflows
Agent operations
2026OpenClaw / LiteLLM / vLLM / Docker / Caddy / TailscalePrivate production operating layer for delegation, browser automation, model fallback, memory, verification gates, and deploy checks.
CRM automation that protects follow-up
CRM automation
2026Next.js / Supabase / webhooks / Vercel / TypeScriptInbound classification, contact upsert, conversation logging, inbox UI, and scheduled follow-up workflow.
Public AI interaction surface
AI workflow demo
2026Next.js / TypeScript / Vercel / AI workflow architectureLive public product surface that demonstrates an AI interaction workflow without exposing private backend controls.
Customer-facing mobile workflow
Mobile product workflow
2026React Native / Expo / TypeScript / Supabase / AI estimatesMobile onboarding, check-ins, chat, food logging, plans, and AI estimate flows combined into one customer-facing workflow.How the work is judged
Workflow, not decorationAI where it earns its place
Models are used for leverage: classification, drafting, routing, enrichment, review, and execution support.
Proof over promise
Screenshots, case studies, live links, build checks, smoke tests, and deploys matter more than strategy decks.
Verified credentials
Public linksReady to evaluate a sprint?
Bring one workflow. Kenneth will map the build path.
Best fit: founders and operators who need one usable workflow shipped, verified, and handed off.