Technical Lead Manager (Physical AI)

TL;DR

Technical Lead Manager (Physical AI): Bridge cutting-edge Machine Learning research and physical robot deployment, leading development and evaluation of Large-Scale Foundation Models (VLAs, World models) for robots and AVs with an accent on model scaling, VLA development, and data strategy. Focus on implementing SOTA architectures, designing robotic-native data pipelines, and translating research into production-ready features.

Location: Work location determines salary and eligibility for benefits (US-based company)

Company

Scale AI is the data engine for the AI industry, providing high-quality data to accelerate AI applications, with the Physical AI team building foundation models for automation in the physical world.

What you will do

• Direct research into scaling laws for Physical AI, utilizing massive datasets for pre-training and fine-tuning generalist policies.
• Develop novel methods for VLA and World model development, including new industry benchmarks.
• Write code to implement, train, and test state-of-the-art architectures; conduct research on data collection, cross-embodiment training, and policy fine-tuning.
• Design robotic-native data pipelines with internal labeling teams, using VLMs for annotation and synthesis.
• Lead and mentor a team of 4-6 Physical AI researchers, fostering high-velocity experimentation.
• Translate research from top conferences into production features and align with cross-functional teams.

Requirements

• Expert proficiency in PyTorch, deep knowledge of Transformer architectures, Attention mechanisms, and Self-Supervised Learning.
• Proven experience with Vision-Language Models (e.g., CLIP, PaLM-E) adapted for spatial reasoning or embodied tasks.
• Experience with Diffusion Models for sequence generation or Generative World Models.
• Strong understanding of Physical AI stack: imitation learning, reinforcement learning, multi-modal sensor fusion.
• Experience with large-scale distributed training on GPU clusters and high-performance data loading.
• 1+ years leading technical teams or projects in research-intensive environments.

Nice to have

• First-author publications at NeurIPS, CVPR, ICRA, CoRL.
• Experience building models generalizing across robot types (arms, mobile bases, humanoids).
• Experience with high-fidelity simulators (Isaac Gym, MuJoCo) and sim-to-real adaptation.

Culture & Benefits

• Comprehensive health, dental, and vision coverage.
• Retirement benefits, learning and development stipend, generous PTO.
• Equity-based compensation subject to approval.
• Potential commuter stipend.
• Inclusive equal opportunity workplace committed to accommodations for disabilities.