Research Scientist Advanced Control

About the role

At Meta IDC (Infrastructure Data Center), the goal is to deliver trusted capacity that powers Meta's AI and products worldwide. The Physical Modeling team develops physics-based and ML models to inform, de-risk, accelerate, and future-proof decisions across the IDC lifecycle.

As Meta's data center fleet grows in scale and complexity, traditional control strategies cannot effectively adapt to transient conditions, multi-system interactions, or next-generation configurations at the pace the fleet scales. The role will build on the team's existing modeling foundation and define and drive a roadmap for intelligent control to improve efficiency, reliability, and sustainability at fleet scale.

Responsibilities

• Define and own the advanced control roadmap in IDC, building on existing physical modeling capabilities.
• Shape the vision for intelligent, autonomous data center operations, from advisory recommendations to governed autonomy at fleet scale.
• Lead projects from problem framing through validated, deployment-ready solutions, translating ambiguous operational challenges into well-scoped research with clear success criteria.
• Develop RL-based control strategies for self-optimizing data center systems, improving thermal stability, energy efficiency, and operational reliability in transient conditions.
• Shape advanced control strategies into deployable solutions aligned with Meta's system architecture, operational constraints, and deployment requirements.
• Establish validation frameworks and safety guardrails that build operational trust.
• Lead cross-functional initiatives involving internal engineering teams and external industrial control system vendors to develop and deliver robust control strategies across Meta's data center fleet.
• Partner with internal engineering teams.

Key facts

• The role is with Meta's Physical Modeling team inside IDC.
• The role defines and owns the advanced control roadmap in IDC.
• Responsibilities include developing RL-based control strategies for data center systems.
• The work includes cross-functional initiatives with internal engineering teams and external industrial control system vendors.