Research Scientist / Engineer - Reasoning

Senior / Staff Research Scientist & Research Engineer — Reasoning and Planning

We are building next-generation intelligent systems capable of operating autonomously in complex, real-world environments. Our team develops large-scale multimodal foundation models, advanced training infrastructure, and integrated systems that enable intelligent decision-making, planning, and long-horizon behavior.

Backed by significant funding and operating at the intersection of AI research, systems engineering, and embodied intelligence, we are investing heavily in next-generation world models and reasoning systems that can generalize beyond static benchmarks into dynamic environments.

We are hiring Research Scientists and Research Engineers across senior to staff levels to advance reasoning and planning capabilities in large multimodal models. This work focuses on enabling systems to decompose goals, reason through multi-step problems, plan actions over long horizons, and adapt in unstructured environments.

What You’ll Do

Research Multi-Step Reasoning and Planning

• Develop methods for reasoning, planning, and decision-making grounded in multimodal world models
• Research architectures and training strategies that improve:
• Compositional generalization
• Long-horizon prediction
• Sequential reasoning
• Multi-step planning

Advance Test-Time Reasoning and Search

• Explore and implement inference-time reasoning methods including:
• Chain-of-thought reasoning
• Beam search
• Monte Carlo Tree Search (MCTS)
• Self-consistency methods
• Verifier-based approaches
• Investigate how model rollouts and search procedures can improve planning and action selection

Improve World Models and Decision-Making

• Study how predictive world models can support inference-time planning and control
• Research methods for integrating reasoning capabilities into large-scale model pipelines
• Collaborate with training and systems teams to scale and deploy research ideas effectively

Build Evaluation and Benchmarking Systems

• Design evaluation benchmarks for reasoning and planning capabilities in dynamic environments
• Develop metrics and testing methodologies for long-horizon task performance and generalization
• Analyze model behavior and failure modes across complex tasks

Collaborate Across Research and Engineering

• Work closely with teams spanning pre-training, post-training, systems, and applied AI
• Translate research ideas into scalable implementations and production-ready systems
• Publish and present research at top-tier conferences and venues

What We’re Looking For

• Strong background in reasoning, planning, search, or sequential decision-making with large models
• Deep understanding of:
• Sequence modeling
• Transformer architectures
• Generative modeling
• Experience with test-time compute and search methods such as:
• Beam search
• MCTS
• Self-consistency
• Verifier-based reasoning
• Strong research judgment and ability to identify high-leverage directions
• Fluency with PyTorch or JAX and ability to implement research ideas end-to-end

Level Expectations

• Senior/MTS candidates are expected to execute complex research projects with strong technical depth and increasing ownership
• Staff-level candidates are expected to define technical direction, shape research strategy, and independently lead major initiatives

Preferred Experience

• PhD in Machine Learning, Robotics, Computer Science, or related fields
• Publications at top-tier venues such as NeurIPS, ICML, ICLR, CoRL, or similar conferences
• Prior work on reasoning in large language or multimodal models
• Experience with:
• Chain-of-thought reasoning
• Process reward models
• Search-based inference methods
• Hierarchical reinforcement learning
• Model-based planning
• Familiarity with:
• Long-horizon prediction
• Video generation
• World model rollouts
• Embodied AI or robotic planning systems

Why This Role Matters

• Work on one of the most challenging open problems in AI: enabling intelligent systems to reason about actions and long-term outcomes
• Research directly impacts real-world behavior and autonomous decision-making
• High degree of research freedom combined with practical deployment opportunities
• Tight collaboration across research, systems, and applied AI teams
• Opportunity to shape the future of large-scale reasoning and planning systems

About the Company

We are a research-driven AI company focused on building scalable intelligent systems capable of robust operation in dynamic environments. By combining advances in machine learning, distributed systems, and large-scale infrastructure, we aim to push the frontier of autonomous intelligence.

We are committed to building an inclusive and diverse workplace and encourage applicants from all backgrounds to apply.