> Markdown version of [/jobs/ext/4519562-senior-staff-machine-learning-engineer-autonomous-driving-foundation-models](https://www.wearedevelopers.com/jobs/ext/4519562-senior-staff-machine-learning-engineer-autonomous-driving-foundation-models). Every page supports `.md` or `Accept: text/markdown`. Links point to the HTML versions so they work for humans too. Agent guide: [/agents.md](https://www.wearedevelopers.com/agents.md).

---

# Senior Staff Machine Learning Engineer - Autonomous Driving Foundation Models

- **Company:** XPeng Motors
- **Location:** Santa Clara, CA, United States
- **Experience:** Expert
- **Salary:** $244,140
- **Contract:** Permanent contract
- **Skills:** Artificial Intelligence, Distributed Computing Environment, Machine Learning, Pytorch, DeepSpeed, Large Language Models, Deep Learning, Data Strategy, Embodied AI, Megatron
- **Published:** November 2, 2026
- **Apply:** https://www.careerbuilder.com/job-details/senior-staff-machine-learning-engineer-autonomous-driving-foundation-models-santa-clara-ca--c93f3cbc-abb7-4818-8e5b-86b6624350b4

## About the Role

Machine Learning, Autonomous Driving Systems, Artificial Intelligence (AI), Autonomous Driving Systems, Machine Learning, Research & Development (R&D), Robotics, Cloning, Architectural Services, Leadership, * 5-7 years of expertise in Deep Learning, with a significant track record in VLM, VLA, or Embodied AI.

 * Proven experience in training and deploying Foundation Models (Transformers, LLMs) at scale.
 * Deep understanding of Sequential Decision Making, World Models, or Policy Gradient methods.
 * Mastery of PyTorch and expertise in distributed training (DeepSpeed, Megatron, etc.).
 * A "Product-First" mindset: The ability to balance cutting-edge research with the deterministic requirements of L4 production vehicles.

## Description

* Architectural Leadership: Lead the design of end-to-end VLA architectures, bridging multi-modal perception with high-level linguistic reasoning and precise action generation.
 * World Model Development: Drive R&D in generative world models (latent dynamics) to create high-fidelity, controllable driving simulations for closed-loop training and evaluation.
 * Policy Evolution: Apply Advanced RL (Online/Offline) and IL to refine driving policies, focusing on long-horizon planning and complex multi-agent interactions.
 * Scaling & Data Strategy: Define scaling laws for driving foundation models, overseeing data curation, automated labeling, and post-training at a multi-billion parameter scale.
 * Global Generalization: Lead the model's adaptation strategy for overseas road conditions, ensuring robust performance across varying traffic laws and driving cultures.