Service
AI Platforms & Orchestration
Durable, multi-model AI pipelines that don't fall over when a provider does.
I build the backend behind AI products: orchestration engines that route work across many models, survive provider failures, and never lose or duplicate a job. I designed and led an AI content-automation platform from scratch - including its workflow engine.
What I can build
Workflow & orchestration engines
DAG-based execution with parallel steps, persistent state and resume-from-checkpoint - jobs continue from the last completed step after a crash instead of restarting from zero.
Multi-model routing & failover
Route each step to the right model across text, image, video and audio; automatic failover and fallbacks when a provider degrades or rate-limits.
Idempotent, reliable delivery
Job-claim locks and idempotent processing so retries and message redeliveries never double-run or duplicate output.
Cost & quality observability
Per-stage latency, inference-quality metrics and alerting to keep an AI pipeline fast, affordable and debuggable.
Proven in production
I built a workflow engine (first Deno, then NestJS) orchestrating 20 models across text, image, video and audio, shipping ~1,200 publications a day to 8 social channels - with automatic failover and durable resume.
See the platform case study →Stack
- Go
- NestJS
- BullMQ
- RabbitMQ
- Redis
- PostgreSQL
- LLM APIs
- Kubernetes
Start a project
Open to contract and advisory work - and full-time roles. Tell me what you're building.