Keshav Pradeep

Currently, I’m an engineer working on scaling Agentic AI at NVIDIA including agent skills, harnesses, multi-agent systems, long-term memory, guardrails, enterprise-scale token savings.

Previously, I completed my undergrad in CS & Math at the University of Maryland, where I researched collective communication in deep learning workloads and network-induced variability in exascale GPU-accelerated supercomputers under Abhinav Bhatele in the Parallel Software and Systems Group.

My research interests include Systems for ML, High-Performance Computing, and AI Agents.

Projects

• NVIDIA/SkillSpector
  Security scanner for AI agent skills that detects vulnerabilities, malicious patterns, and security risks before installation. 4k+ GitHub stars.

Conference Papers

• The Big Send-off: Scalable and Performant Collectives for Deep Learning
  Siddharth Singh, Keshav Pradeep, Mahua Singh, Cunyang Wei, Abhinav Bhatele
  IPDPS 2026 · pdf · website

• The Case of the Elusive Application Performance on Production GPU Supercomputers
  Cunyang Wei, Keshav Pradeep, Abhinav Bhatele
  IPDPS 2026 · pdf · website

Posters

• Optimizing Collectives with Large Payloads on GPU-based Supercomputers
  Siddharth Singh, Mahua Singh, Keshav Pradeep, Abhinav Bhatele
  SC 2025 · poster · abstract

• Unmasking Performance Variability in GPU Codes on Production Supercomputers
  Cunyang Wei, Keshav Pradeep, Abhinav Bhatele
  SC 2025 · poster · abstract

• Unmasking Performance Variability in GPU Codes on Production Supercomputers
  Cunyang Wei, Keshav Pradeep, Abhinav Bhatele
  MUG 2025 · poster