AI Research Scientist, Visual Coding - MSL FAIR

Meta; Menlo Park (CA); On-site; USD 180,000 - 280,000; Full time

Job summary

Meta is seeking AI research scientists to advance multimodal large language models and related data pipelines. You will work across pre-training, mid-training, and post-training stages, tackling challenges in image, video, audio, and 3D modalities.

You will lead major technical initiatives with cross-functional teams, publish influential research, and contribute to visual code generation and data curation at scale within the FAIR/MSL ecosystem.

Qualifications

• PhD or equivalent research credentials in Computer Science or related field.
• 2+ years industry research experience with large language or multimodal models.
• Experience leading major technical initiatives with cross-functional impact.
• Proficiency in Python and PyTorch; experience with Spark or Ray.
• First-author publications in leading peer-reviewed conferences (ACL, NeurIPS, ICML, ICLR, AAAI, KDD, CVPR, ICCV).

Responsibilities

• Contribute to Meta's frontier multimodal models across pre-training, mid-training and post-training stages.
• Lead technical initiatives with cross-functional impact and influence strategy across teams.
• Collaborate with scientists and engineers across Meta's FAIR and MSL groups.
• Advance visual coding capabilities and data curation at scale for training and evaluation.

Skills

Research leadership

Multimodal AI/LLM experience

Python programming

Cross-functional collaboration

Publications in top conferences

Data curation at scale

Experimentation and analysis

Systems understanding of large models

Education

PhD in Computer Science or related field

Bachelor's degree in CS/CE or related field

Tools

PyTorch

Spark

Ray

DataFlow

Hive

Parquet

Iceberg

Job description

We are seeking AI research scientists to help us build Meta's most advanced multimodal large language models. To push the frontier of what those models can generate, we need to work with data and train at scale, and advance our visual coding capabilities.

Our team contributes across all stages of development (pre-training, mid-training, post-training) and all multimodal modalities (image, video, audio, 3D). We are tackling complex challenges at trillion-scale, including organic data curation, synthetic data generation, agent and interaction data, visual code generation, agent post-training, and frontier paradigms that redefine what is possible.

Based in Meta Superintelligence Labs (MSL) within the Fundamental AI Research Organization (FAIR), you will directly contribute to Meta's frontier models and collaborate with scientists and engineers across MSL.

• Bachelor's degree in Computer Science, Computer Engineering, relevant technical field, or equivalent practical experience
• Bachelor's degree in Computer Science, Computer Engineering, a relevant technical field, or equivalent practical experience
• A PhD in Computer Science or a related technical field
• 2+ years of industry research experience with large language or large media models, computer vision, or related AI/ML models
• Experience as a formal technical lead, leading major technical initiatives with cross-functional impact, and/or influencing strategy across multiple teams
• Experience in visual code generation - generating code that produces visual output for 3D, web, or game environments
• Experience with multimodal pre-training or mid-training data curation for large media perception models
• Published research in leading peer-reviewed conferences (e.g., ACL, NeurIPS, ICML, ICLR, AAAI, KDD, CVPR, ICCV) and/or demonstrated significant industry influence in the field of AI Experience working on frontier-quality and state-of-the-art Large Language or Large Media Models
• Programming experience in Python and hands-on experience with frameworks like PyTorch or Spark, or related distributed computing frameworks (Ray, DataFlow)
• Familiarity with SQL and file formats, such as Hive, Iceberg, Parquet, etc
• First-author publications in leading peer-reviewed conferences (e.g., ACL, NeurIPS, ICML, ICLR, AAAI, KDD, CVPR, ICCV)