Research Scientist, Infrastructure Modeling and Reliability

Job description

Meta builds technologies that help people connect, find communities, and grow businesses. Meta's infrastructure supports services used by billions of people, and operating that infrastructure efficiently requires increasingly sophisticated modeling of demand, utilization, reliability, and physical resource constraints. We are seeking an experienced Research Scientist or Applied Scientist to define and build new modeling approaches for power utilization across Meta's infrastructure.

This role will lead the development of statistical and machine learning models that monitor power consumption, project peak demand, quantify uncertainty, and inform how Meta maximizes usable power within failure domains while maintaining target reliability levels. Qualified candidates will have has deep experience modeling high-dimensional, noisy, and interdependent systems, and has demonstrated the ability to translate scientific advances into production systems that influence large-scale infrastructure strategy.

Responsibilities

• Define the scientific and technical strategy for modeling power consumption, peak risk, and reliability tradeoffs across large-scale infrastructure systems
• Develop statistical, machine learning, and/or optimization models that forecast power demand, estimate peak distributions, quantify uncertainty, and support operational decision-making
• Build approaches that reason about high-dimensional signals, correlated demand, failure-domain constraints, reserve margins, and reliability targets
• Partner with engineering, capacity planning, data center, energy, hardware, operations, and finance teams to translate model outputs into infrastructure planning and utilization decisions
• Establish evaluation frameworks, backtesting methods, confidence intervals, and monitoring systems to measure model quality and operational risk
• Identify opportunities to safely increase power utilization, reduce stranded capacity, improve cost efficiency, and guide long-term infrastructure investment
• Lead ambiguous, company-critical technical initiatives across organizations, influencing strategy and aligning stakeholders around scientifically grounded decisions
• Mentor other scientists and engineers, raise the technical bar for modeling and forecasting systems, and represent Meta's work through appropriate external publications, talks, or industry engagement

Qualifications

• 10+ years of experience developing statistical, machine learning, simulation, forecasting, optimization, or other quantitative modeling systems Experience leading ambiguous, cross-functional technical programs from problem definition through model development, evaluation, deployment, and business impact Experience coding in Python, R, C++, Java, or similar languages for data analysis, modeling, simulation, or production systems Experience communicating complex technical concepts, assumptions, uncertainty, and tradeoffs to technical and non-technical audiences Experience influencing technical strategy across multiple teams or organizations
• Bachelor's degree in Computer Science, Computer Engineering, relevant technical field, or equivalent practical experience Experience with infrastructure, capacity planning, power systems, energy systems, data centers, reliability engineering, distributed systems, supply-chain optimization, or resource allocation Experience developing peak-demand forecasts, confidence intervals, risk estimates, anomaly detection, or backtesting frameworks Experience mentoring senior technical contributors and building scientific communities across organizations
• Demonstrated record of industry-level technical leadership, such as defining new research directions, influencing company strategy, publishing in leading venues, or shaping external technical standards Experience building models that support operational decisions under explicit reliability, safety, cost, or utilization constraints Experience applying optimization, operations research, or decision science to large-scale resource planning Experience modeling high-dimensional, sparse, noisy, or strongly correlated data in production environments Experience with time-series forecasting, probabilistic forecasting, Bayesian modeling, extreme-value modeling, causal inference, stochastic processes, simulation, or uncertainty quantification