# Raphael Mansueto

Raphael Mansueto is a Senior Full-Stack Engineer specializing in AI integrations, based in Cebu, Philippines. His official current role is Senior Full Stack Developer at VISEO. Previously, he worked as a Full-Stack AI Integration Engineer at HustleWing, combining product delivery, Go services, asynchronous AI processing, and production automation in one employment chapter. He builds reliable TypeScript and Go systems across web and mobile.

## Primary pages

- Portfolio: https://raphaelmansueto.com/
- Ample News case study: https://raphaelmansueto.com/work/ample-news
- KudosCourts case study: https://raphaelmansueto.com/work/kudoscourts
- CravingsPH case study: https://raphaelmansueto.com/work/cravingsph
- Engineering notes: https://raphaelmansueto.com/engineering
- Human decision gates in production AI workflows: https://raphaelmansueto.com/engineering/human-decision-gates-production-ai
- Tracing AI workflows across the whole product: https://raphaelmansueto.com/engineering/trace-ai-workflows-end-to-end
- Transactional boundaries for competing reservations: https://raphaelmansueto.com/engineering/transactional-reservation-boundaries
- Realtime reconciliation with React Query: https://raphaelmansueto.com/engineering/realtime-react-query-reconciliation
- Resume: https://raphaelmansueto.com/resume.pdf

## Verified areas of work

- Production AI workflows with structured output, human review, orchestration, and observability
- Full-stack systems with Next.js, React, Node.js, Go, PostgreSQL, and typed API contracts
- Web and mobile delivery with Next.js, Expo, and React Native
- Integration work across queues, webhooks, media providers, payments, EVM, and Solana
- Reliability patterns including state machines, idempotency, row locking, outbox delivery, and validation

## Featured work

- Ample News: A production AI workflow across grounded research, agents, voice, video, social publishing, and observability. Evidence: Long-running work stays reviewable and recoverable: Each stage records progress before work moves across people or providers. One run identity connects the operating story: Correlated signals make model behavior inspectable inside the surrounding product workflow.
- KudosCourts: Transactional reservation commands, lifecycle events, and targeted realtime reconciliation across web and mobile. Evidence: One accepted booking, every affected view in sync: The booking decision commits before lightweight events tell affected screens what to refresh.
- CravingsPH: Transaction design across three-region order state, deterministic locking, realtime reconciliation, and kiosk trust. Evidence: Concurrent work resolves to one trusted state: A guarded command commits coordinated restaurant state before realtime screens reconcile.

## Engineering notes

- Human decision gates in production AI workflows: Separate generation from mutation. Let the system gather context and prepare a typed proposal, then require an explicit human decision before application code performs a state-changing action.
- Tracing AI workflows across the whole product: Carry one run identity across model calls, workflow stages, tools, application logs, and provider callbacks. Each telemetry system can keep its specialty while the shared identity reconstructs one product outcome.
- Transactional boundaries for competing reservations: Treat pending requests as intent, then make acceptance the inventory-claiming transaction. Lock the reservation target first, recheck availability, commit one winner, and close overlapping requests coherently.
- Realtime reconciliation with React Query: Publish a minimal lifecycle event after commit, invalidate only the affected React Query keys, and refetch canonical state. Realtime improves responsiveness without owning correctness.

## Professional profiles

- GitHub: https://github.com/raphaelmans
- LinkedIn: https://linkedin.com/in/raphaelmansueto
- Email: mailto:raphaelmansueto@gmail.com

Last reviewed: 2026-08-12