Fengshuang Tan

I’m a Senior Software Engineer with 10+ years at Red Hat, delivering AI Agent/LLM systems end-to-end—from architecture to production operations. I own building blocks like multi-agent RCA frameworks with LangGraph/AutoGen and RAG knowledge assistants that teams can use for 24/7 self-service.

I’m especially strong in turning messy distributed-system failures into automated, self-service tooling. I integrated LLM-based log analysis into Prow CI pipelines and added an LLM-as-judge evaluation layer so low-confidence RCA results are flagged for human review.

On the infrastructure side, I design for reliability and scalability: distributed Redis-queue pipelines, AWS EKS deployments with auto-scaling worker pools, Kubernetes readiness/liveness probes, and safe rollouts with rollback on failed health checks. I also operate CI/CD platforms on OpenShift with 100% config-as-code (JCasC), secrets management, and production-grade observability.

Beyond core delivery, I’ve led integration testing for OpenShift Platform Plus, diagnosed chronic platform instability, and improved stability from 95%+ while reducing cross-team resolution cycles. I bring a practical engineering ethos—secure, observable, test-gated workflows that reduce time-to-fix and keep production signals clean.