is a systems researcher,
engineer, and
builder.
I'm a graduate student in Computer Science at Georgia Tech, advised by Prof. Alexey Tumanov in the Systems for AI Lab.
I build production systems that balance performance, reliability, and research rigor across inference, serverless, and hybrid cloud workloads.
Most recently I interned with AWS DynamoDB, where I designed leadership failover tooling for JournalDB. Earlier, I led cross-cloud orchestration efforts at the Distributed Systems Lab, IISc.
Recent projects:
distributed master-worker topology with fault tolerance
ILP-guided scheduling for mixed GPU fleets
adaptive branching policies for inference serving
My research focuses on systems for AI, with an emphasis on programmable inference serving, distributed scheduling, and cloud-native control planes. I draw from compiler design, large-scale observability, and hardware-aware optimization to build dependable platforms for emerging AI workloads.
At the Systems for AI Lab, I'm designing policy-decoupled C++ abstractions for programmable inference serving that separate deployment topology from scheduling logic—delivering 12% accuracy gains, 99.99% SLO attainment, and 20K QPS throughput on heterogeneous clusters.
I previously shipped XFaaS, a cross-cloud orchestration platform that delivered 75% latency reductions and 57% cost savings through adaptive placement with telemetry-informed routing, collaborating with industry partners to operationalize hybrid quantum-classical workflows.
I'm also the Head TA for ECE 4150 — Cloud Computing at Georgia Tech (Spring 2025), leading course delivery, AWS labs, and mentoring cloud-native capstones.
You can find me on LinkedIn, GitHub, or Google Scholar. Feel free to reach out at tkhare7@gatech.edu.
