What I work on
Self-hosted LLMs, RAG pipelines, inference routing, multimodal systems (ASR/TTS/vision). Privacy-first, model-agnostic architectures.
Kubernetes, multi-tenant isolation, workflow orchestration, CI/CD, secrets management, observability at scale.
Event-driven architectures, message queues, database scaling, high-availability systems, and production reliability.
Experience
Building a deterministic workflow orchestration engine in Rust with precompiled graphs for low-latency execution. Designing self-hosted, model-agnostic AI runtime with confidence-based routing.
Led architecture of Stateset Cloud, a Kubernetes-based multi-tenant platform for self-service deployment of backend services, workflows, and AI workloads.
Architected AI-driven geospatial address correction pipeline. Led early Kubernetes adoption and cloud-native architecture initiatives across production systems.