Nathan Zelalem Kebede

I’m a Software Engineer who enjoys building production-ready systems that reduce operational risk while improving performance and user impact. Most recently, I worked on the S3 team at Amazon WebServices (AWS), focusing on safer deployments and faster incident response.

At AWS S3, I helped build a pre-production stage for the S3 Express One zonemetering workflow to validate changes before rollout. I reduced operational overhead by designing and implementing an Object Cleanup Lambda, and I built a debug mode with structured logging and an operational runbook to reduce RCA time during incidents.

Earlier, at Africa to Silicon Valley (A2SV), I collaborated with frontend and mobile teams to develop APIs using Python, FastAPI, PostgreSQL, and REST—growing the user base from 110 to 1,100 and messages from 3,000 to 30,000 in seven months. I also integrated LLMs and image generation models into AfroChat, and implemented RAG with LangChain and Unstructured, improving retrieval efficiency from 25% to 80%.

On my current focus, I’m building an RDBMS-backed message queue with Kotlin, Spring Boot 3, PostgreSQL, Docker, and Terraform on GCP. I’ve designed for high throughput (296% gain, 57% P95 reduction), handled real reliability issues like a Linux OOM kill and connection problems, and instrumented the service with Micrometer/OpenTelemetry for measurable performance and alerting.