Mehul Vani

About Mehul Vani

I build scalable backend systems and distributed services that power production cloud platforms. My work focuses on designing reliable APIs, improving efficiency, and enabling safe operations across large multi-region environments.At Oracle Cloud Infrastructure, I work on backend services that support compute and AI workloads in Kubernetes-based platforms, including orchestration APIs, CI/CD automation, observability pipelines, and self-healing mechanisms to reduce manual intervention and improve availability.Previously at Amazon, I built fault-tolerant microservices handling high-volume payments workflows (charges, refunds, disbursements) and worked on improving consistency, reconciliation, latency, and throughput across customer-critical paths.I enjoy solving hard distributed systems problems—especially reliability, automation, and performance at scale and I’m looking for backend roles where production responsibility and technical ownership matter.

Experience

1. Member of Technical Staff (Sde 2)

   Oracle

   Aug 2024 — Present · Seattle, WA, US

   Designed and optimized distributed backend systems and services for Kubernetes-based compute workloads, improving reliability by 35% and reducing latency by 25%(Java/Python/PostgreSQL).• Built a backend platform for LLM orchestration (vLLM inference, RAG pipelines, vector embeddings, OCI Vector Search, classifiers, fine-tuning) enabling scalable AI assistant capabilities.• Engineered compute orchestration APIs for AAIL enabling cross-architecture deployments (x86/GPU/ARM) and non-destructive node cycling (zero-downtime upgrades), reducing cost by 20%.• Automated CI/CD and release workflows with Terraform and Go (rollback hooks, observability checks), cutting manual release effort by 50%, and delivered OpenTelemetry collectors + Elasticsearch indexing and Python/Flask on-call automation to improve visibility and cut reporting time from 3 days to hours.

Education

• Binghamton University

  Master's degree, Computer Science