Nishant Gupta is a Staff Software Engineer and Researcher at Meta, where he specializes in building large-scale distributed systems and AI infrastructure that power production-scale intelligent systems. His work sits at the intersection of distributed computing, systems architecture, and agentic AI, with a focus on designing reliable, scalable, and operationally resilient infrastructure for autonomous software systems.

Within Meta Superintelligence Labs (MSL), Nishant works on agentic infrastructure - platforms that enable AI agents to operate safely and effectively within complex distributed environments. His work focuses on evaluation and auditing of AI-driven decision systems, reliability and safety for autonomous workflows, human oversight mechanisms, feedback-driven system optimization, and measuring real-world production impact beyond offline benchmarks.

Prior to his work on AI infrastructure, Nishant led the development of Meta's next-generation elastic compute infrastructure, helping manage approximately 30% of Meta's global compute capacity across tens of millions of servers spanning around 20 geo-distributed data centers. His work significantly improved fleet-wide utilization while maintaining strict service-level objectives, contributing to infrastructure efficiency at one of the world's largest computing platforms. He also partnered with engineering leadership across Ads, Instagram, WhatsApp, Finance, and Infrastructure to define multi-year strategies for improving global compute efficiency.

His research focuses on bridging theoretical advances with production engineering. He is the lead author of Dynamic Idle Resource Leasing to Safely Oversubscribe Capacity at Scale, which introduced a production system for leasing idle datacenter capacity while preserving reliability and strict SLO guarantees through rigorous simulation, controlled experimentation, and real-world safety validation. His publications have received more than 90 scholarly citations.

Nishant is passionate about designing systems that remain reliable under uncertainty. His work explores how distributed services and AI agents coordinate, communicate, and make decisions while maintaining correctness, observability, fault tolerance, and operational safety. Throughout his career, he has led initiatives that:

Architect highly available distributed platforms operating at massive scale Build production-ready agentic infrastructure for autonomous AI systems Design resilient systems with strong observability, graceful degradation, and operational reliability Translate cutting-edge research into practical engineering patterns adopted in production

He is an active international speaker on AI infrastructure, distributed systems, production AI, and agentic engineering, regularly presenting at leading developer conferences and industry events.

Nishant earned his Master of Science in Computer Science from the University of California, Los Angeles (UCLA), where his research focused on scalable distributed systems and machine learning.

Collaboration: - Keynotes and conference presentations - Technical workshops and executive briefings - Advisory engagements on AI infrastructure, distributed systems, and production AI - Partnerships with AI platforms, developer tools, and engineering organizations

Contact: nishantgupta@g.ucla.edu