Haystack
← Back to Jobs
Other
PG

Physics Simulation & Generative AI Research Scientist

PROLIM Global CorporationUnited States🇺🇸United StatesPosted Oct 8, 2026

Quick Overview

Seniority
Mid Senior
Work mode
Hybrid
Location
United States
Posted
Yesterday
RoboticsERPGenerative AIPyTorchPython

Job Description

PROLIM Global Corporation ( is hiring a Research Scientist – Physics Simulation & Generative AI for a leading Technology customer.
Research Scientist – Physics Simulation & Generative AI
Location:
Remote – US
Job Description

We are seeking a highly experienced Research Scientist V to work on physics-based motion rewards used to post-train generative video models through reinforcement learning (RL).

The researcher will focus on improving physics-based motion evaluation, optimizing simulation pipelines for large-scale RL training, and extending motion modeling from single-body plausibility to human-object and human-scene interactions using 3D mesh representations.
Key Responsibilities

·         Improve dynamic and contact terms within physics-based motion reward signals.

·         Optimize physics-based rewards to run efficiently inside large-scale RL training loops.

·         Extend motion evaluation from single-body motion to human-object and human-scene interactions.

·         Develop and work with 3D human pose, motion, and mesh representations.

·         Implement and optimize compute-intensive simulation and ML pipelines using Python and PyTorch.

·         Apply numerical methods and ensure simulation/solver stability.

·         Collaborate on post-training and reinforcement learning approaches for generative video models.

·         Validate simulated physical quantities against human-judged ground truth when applicable.

 

Required Skills

·         5+ years of relevant research/industry experience.

·         Strong experience with rigid-body and contact simulation using MuJoCo, Isaac, Bullet, PhysX, or equivalent.

·         Experience with 3D human pose and motion representations, preferably SMPL/SMPL-family models.

·         Advanced Python and PyTorch skills.

·         Experience profiling and optimizing compute-intensive ML or simulation pipelines.

·         Experience with reinforcement learning reward design and/or post-training of large generative models.

·         Strong knowledge of numerical methods and solver stability.

·         Experience with physics-based simulation and motion modeling.

 

Preferred / Nice-to-Have Skills

·         Generative video or diffusion model experience.

·         Publications in physics-grounded generative modeling or differentiable simulation.

·         Experience scaling simulation workloads.

·         Experience validating simulated physical quantities against human-judged ground truth.

·         Experience with human-object or human-scene interaction modeling.

 

Education

Ph.D. preferred in Computer Science, Robotics, Graphics, Mechanical Engineering, Physical Science, or a related field. Equivalent industry research experience may be considered.


Interested Candidates

Please send your updated resume and contact details to:

 

About PROLIM Corporation

PROLIM is a leading provider of end-to-end IT, PLM, and Engineering Services and Solutions for Global 1000 companies. We understand business as much as technology and help our customers improve profitability and efficiency through high-value consulting, staffing, engineering, and project management services.
Our offerings include Advisory Services, PLM Software & Services, Program Management, Solution Architecture, Training & Staffing, Cloud Solutions, Infrastructure Services, ERP Practices, Quality Assurance Services, and Engineering Services across a wide range of industries.

Similar jobs