Also building AI-driven systems — architected an internal multi-agent platform (LangGraph, FastAPI, Celery) that automates PR risk analysis and code remediation using LLMs (OpenAI/Ollama), cutting manual review time by ~35%.

Most of my work lives at the intersection of distributed systems and observability — designing for failure, reducing MTTD, and making production less surprising.

Currently at Zinnia, leading backend development on a correspondence platform processing millions of insurance documents.

Core

APIs

Databases

Message Streaming

Cloud & DevOps

Observability & Reliability

• kubernetes-sigs/kubebuilder — Added support for sidecar containers, multiple Deployments, and multiple ServiceAccounts in Helm chart generation (open)
• kubernetes-sigs/kueue — Authored KEP-13502: configurable timeout for pods stuck unscheduled post-admission, implemented across Job, Pod, JobSet, Ray*, MPIJob, and 4+ other job integrations (open)
• k9s — Fixed an input-handling issue in the Kubernetes CLI tool. Random characters were being input to k9s after periods of inactivity.
• Istio — Diagnosed and fixed a Gateway API HTTPRoute bug where RegularExpression routes were ranked below PathPrefix routes in Envoy config generation, causing incorrect request routing (closed, pending maintainer feedback)

• Expanding expertise in SRE and DevOps practices — automation, observability, and infrastructure as code.
• Designing scalable backend systems in Go and Java, integrating reliability engineering principles from day one.
• Strengthening hands-on proficiency with Terraform, Kubernetes, and AWS for high-availability architectures.
• Writing about Kafka internals, Go production patterns, and Kubernetes failure modes on LinkedIn