About the role#
XPENG is seeking an intern to support the optimization and deployment of Vision-Language-Action (VLA) models onto vehicle-grade compute platforms. This role focuses on bridging the gap between research models and real-time, production-ready systems for autonomous driving applications.
What you'll do#
- Support model quantization and deployment efforts for large-scale multimodal models, including Transformers and vision-language models.
- Assist with applying model optimization techniques such as post-training quantization, quantization-aware training, pruning, and related compression methods under the guidance of senior engineers.
- Collaborate with research and platform teams to improve model deployability and address hardware and runtime constraints.
- Contribute to the development of deployment tools, test pipelines, and runtime modules using C++ and Python.
- Analyze model performance, including memory usage, latency, and numerical accuracy across various deployment targets.
- Participate in debugging and performance tuning across the model, runtime, and system stack, while supporting validation workflows for vehicle and simulation environments.
What you'll need#
- Currently pursuing a BS, MS, or PhD in Computer Science, Electrical Engineering, Robotics, or a related field.
- Strong programming skills in C++ and/or Python.
- Familiarity with deep learning frameworks, specifically PyTorch.
- Basic understanding of model inference, deployment, or optimization workflows using tools such as ONNX or TensorRT.
- Exposure to model compression or quantization concepts like INT8 or FP16.
- Interest in computer architecture, performance optimization, and edge or embedded systems.
Location & details#
- Location: Santa Clara, California
- Term: Summer 2026
- Modality: On-site
- Employment Type: Full-time
- Compensation: Paid
