Hi, I'm @aravindsundaresan,
Aravind Sundaresan
Distributed Systems & ML Infrastructure Engineer
Microsoft • Ex-Amazon
Building scalable inference infrastructure, GPU scheduling systems, and Kubernetes-native orchestration platforms.
// Focus
I'm an infrastructure engineer with 7+ years of end-to-end ownership — designing, launching, and maintaining large-scale distributed backends. At Microsoft, I built a metadata platform serving 17,000+ microservices — Redis-based adaptive concurrency control and active-passive multi-region failover — and, separately, a C++ Test-in-Production telemetry framework adopted across 7–10 engineering teams. At Amazon, I built an automated hardware scheduler fleet for 50+ physical devices designed to self-heal via ADB-based automatic recovery.
Since late 2023, my core direction has focused on ML systems and inference performance. I conduct active systems research to mitigate throughput bottlenecks in modern serving engines, shipping open-source systems research (Clairvoyant, ACO, ServiceScope) spanning inference scheduling, GPU cluster orchestration, and pre-deployment dependency mapping.
[ SYSTEMS TOOLING DIRECTORY ]
// Systems Languages
// Orchestration & Compute
// ML Serving Runtimes
// Distributed Data & Core
// Featured Systems Projects
// Technical Writing
// Work Experience
// Current Research Interests
// Let's talk
Focused on inference infrastructure, GPU scheduling, distributed systems, and ML-native orchestration.
Open to ML systems, platform infrastructure, and distributed compute engineering roles globally. Based in Hyderabad and open to relocation.