AI Research Scientist, MRS Knowledge

The MRS AI Knowledge team is at the forefront of innovation in content understanding, user understanding, retrieval, and ranking. As a key driver of Meta's app growth, we are committed to revolutionizing recommendation systems by leveraging the latest advancements in AI, including large language models (LLMs) and generative AI. We're dedicated to pushing the boundaries of recommendation systems to deliver user experiences across Facebook, Instagram, Threads, and more.

Responsibilities

• Design and execute original research in core machine learning, including areas such as deep learning architectures, optimization methods, representation learning, or generative modeling
• Develop and maintain high-quality, reusable research code and experimental pipelines to support reproducible and scalable ML research
• Identify and define long-term research directions as well as near-term milestones that align with team and organizational goals
• Analyze experimental results using rigorous statistical and empirical methods to draw actionable conclusions and inform future research directions
• Contribute to publications at top-tier machine learning venues and support open-sourcing efforts that advance the broader research community
• Collaborate with product and engineering teams to translate fundamental research findings into practical improvements in Meta's AI systems
• Establish and improve telemetry, evaluation frameworks, and benchmarks to measure research progress and model quality
• Provide technical guidance and mentorship to other engineers and researchers on the team, sharing expertise through design reviews, research critiques, and knowledge-sharing sessions
• Drive cross-functional research initiatives by coordinating with collaborators across AI infrastructure, product, and applied science teams
• Proactively identify gaps in existing research programs or methodologies and propose improvements to strengthen the team's research output

Minimum qualifications

• Bachelor's degree in Computer Science, Computer Engineering, relevant technical field, or equivalent practical experience
• 6+ years of experience in machine learning research, including hands-on work with deep learning, optimization, generative models, or related core ML areas
• Experience developing and publishing novel machine learning approaches at peer-reviewed venues such as NeurIPS, ICML, ICLR, CVPR, or equivalent
• Experience designing and running large-scale ML experiments, including model training, evaluation, and iterative refinement using frameworks such as PyTorch or TensorFlow
• Experience communicating research findings and technical trade-offs in writing to both research and non-research audiences, such as through papers, technical reports, or design documents
• Experience collaborating across research and engineering teams to drive projects from ideation through production impact Experience mentoring other researchers or engineers and contributing to the development of team research culture, evaluation standards, or shared tooling
• Experience applying core ML research to large-scale production systems, including working with high-dimensional, large-volume datasets and distributed training infrastructure
• PhD in Machine Learning, Computer Science, Statistics, or a closely related field, with a dissertation focused on a core ML topic
• Proven track record of high-impact publications or open-source contributions that have been widely adopted or cited within the ML research community

Preferred qualifications

• Experience mentoring other researchers or engineers and contributing to the development of team research culture, evaluation standards, or shared tooling
• Experience applying core ML research to large-scale production systems, including working with high-dimensional, large-volume datasets and distributed training infrastructure
• PhD in Machine Learning, Computer Science, Statistics, or a closely related field, with a dissertation focused on a core ML topic
• Proven track record of high-impact publications or open-source contributions that have been widely adopted or cited within the ML research community