Product Engineer · AI Systems · Distributed Architecture
Based in Paris, France. I design and ship production-grade systems end to end — from architecture to deployment. Backend-heavy, fullstack in practice, product-minded in approach.
Currently building infrastructure where AI pipelines and real-world data meet.
Systems that process intent and deliver structured, actionable outputs at scale. Multi-stage LLM pipelines. Async distributed architectures. Real-time event processing. APIs that other teams adopt and keep running long after handover.
Not prototypes. Production.
Natural language query → structured, enriched business data in under 30 seconds
Built entirely from 0 → production as Founding Engineer. The system runs a 4-stage LLM pipeline: intent classification, semantic category matching via embeddings (TF-IDF fallback), external search, and async enrichment with 200 parallel workers. Redis handles queues, job state, caching, and rate limiting. Deployed on GCP Cloud Run with full CI/CD, observability, and per-query cost tracking.
Stack: Python · FastAPI · asyncio · NestJS · · Laravel · PHP · TypeScript · Redis · PostgreSQL · GCP Cloud Run · Docker · GitHub Actions
100,000+ requests/hour · fraud detection · automated alerting
Architected end-to-end with the CTO and CEO. The system ingests events from all client platforms in real time, applies rule-based threshold logic (GGR limits, win anomaly detection, irregular activity patterns), caches computed metrics via Redis to eliminate recalculation, and fires automated multi-channel alerts on breach. Directly reduced fraudulent activity and financial losses across platforms.
Stack: Node.js · Bun · Redis · PostgreSQL · event-driven architecture
Backend Node.js · TypeScript · Python · FastAPI · NestJS · Express.js · PHP · Laravel
Frontend React · Next.js · Vue · Nuxt
Cloud GCP (Cloud Run, Cloud SQL, Logging, Monitoring) · AWS (Lambda, ECS, SNS, SES) · Azure · Docker
Data PostgreSQL · MySQL · MongoDB · Redis (cache · queues · jobs)
Observability Grafana · Prometheus · CloudWatch · GCP Cloud Monitoring
AI LLM integration · semantic embeddings · async inference pipelines
CI/CD GitHub Actions · multi-env deployment (dev → UAT → prod)
Security JWT · HTTP-only cookies · rate limiting · CORS · CSP · circuit breaker
AI tools Cursor · Claude Code · Codex
| Category | Providers |
|---|---|
| Payment gateways | Stripe · PayPal · Mollie · Systempay · Paystack · Flutterwave · Ecocash · Mpesa · PayLesotho |
| CRM platforms | Netcore · FastTrack |
| Cloud messaging | AWS SNS · AWS SES · Azure ACS · SendGrid |
| Calendar APIs | Google Calendar · Outlook |
| Auth | Google OAuth2 · Facebook OAuth2 · Clerk |
| AI APIs | Google Gemini · Grok · OpenAI · AI inference APIs |
| Email / marketing | Postmark · Loops · Sendgrid |
- Building agentic data infrastructure at the intersection of LLMs and real-world business data
- Expanding my knowledge on AI and contributing to open-source projects
- Open to founding engineer and product engineer roles in Paris or remote
I write about systems I've actually built — not tutorials.
- Never Miss a Cloud Run Error: A Guide to Comprehensive GCP Alerting
- Why NestJS Hot Reload Does Not Work in Docker and How to Fix It Properly
- Using TypeORM with TSX: A Smoother Development Experience
- More on dev.to →
Check out my latest posts here: dev.to/osalumense
