Senior Machine Learning Engineer - Generative Models

Applied Intuition•Sunnyvale, CA

2d•$185,000 - $260,000•Onsite

About The Position

Applied Intuition, Inc. is seeking senior machine learning engineers to advance the generative modeling technology behind Neural Simulation, their state-of-the-art product for turning real-world driving data into high-fidelity, photorealistic simulation environments. This role involves pushing the boundaries of diffusion and video generation models to create realistic, controllable sensor data, augment real-world logs with new scenarios, and enhance the product's utility for customers training and validating autonomy systems. The work will directly influence the development of next-generation data-driven autonomous vehicles for major OEMs. This position is ideal for engineers skilled in generative modeling, computer vision, and machine learning, who are eager to apply the latest research to production challenges in Physical AI simulation.

Requirements

5+ years of experience developing and shipping ML or computer vision systems
A minimum of a Bachelor's degree in computer science, physics, robotics, or equivalent
Strong hands-on experience with diffusion models and video generation (e.g., latent and video diffusion models, diffusion transformers)
A solid foundation in generative modeling and deep learning, including training and fine-tuning large models
Proficiency in Python and PyTorch
A proven ability to turn research ideas into robust, production-quality software
Strong problem-solving skills and comfort with ambiguity

Nice To Haves

Experience with learning-based 3D reconstruction, such as 3D Gaussian Splatting, NeRFs, or feed-forward Gaussian Splatting
A background in computer vision (e.g., SfM, SLAM, photogrammetry) and/or computer graphics (e.g., rendering, rasterization, ray tracing)
A track record of shipping ML products with clearly defined evaluation metrics and benchmarks
Experience in autonomous driving or robotics, including working with multi-sensor data (camera, LiDAR, radar)
Experience with 3D-aware or multi-view consistent generation, or world models
Experience with large-scale distributed training and inference optimization (e.g., distillation, efficient sampling)
Programming experience in C++ and/or CUDA
Peer-reviewed research at conferences such as CVPR, ICCV/ECCV, NeurIPS, ICLR, ICML, or SIGGRAPH
A Master's degree or PhD in computer science, physics, robotics, or related fields

Responsibilities

Develop and advance diffusion and video generation models that power our Neural Simulation product, bringing the latest research into production to improve realism, controllability, and scalability
Push the limits of generative simulation for driving scenes, including controllable generation conditioned on scene layout, camera pose, actors, and trajectories
Develop temporally consistent, multi-camera video generation
Augment real-world logs with new scenarios, actors, and conditions such as weather and lighting
Combine generative models with our neural reconstruction pipeline to improve fidelity and coverage of simulated scenes
Scale training and inference of large generative models for production workloads
Define and build evaluation metrics, benchmarks, and validation workflows that measure realism, temporal consistency, controllability, and sim-to-real gap
Work closely with customers to understand their pain points and implement technical solutions in the Neural Simulation product
Collaborate closely with Infra, Autonomy, Research and other product teams to deliver end-to-end solutions
Take ownership of critical technical components and influence architecture and product decisions