Research Intern, Scene Understanding and Planning

DiDi; San Jose (CA); On-site; USD 52,000 - 74,000; Full time

Job summary

DiDi’s Autonomous Driving unit is seeking a Research Intern to tackle open problems in perception, prediction, and planning for real-world driving. You will work on hard tasks from problem definition to validated solutions, often with publishable results, under the guidance of team leads.

You will design deep learning models for real-time scene understanding, train on driving logs and public benchmarks, and collaborate with cross-functional teams to transition prototypes into production-ready

Qualifications

• Currently enrolled in an MS or PhD program in Computer Science, Robotics, Electrical Engineering, Machine Learning, or a related field.
• Available for a full-time onsite internship of 3–6 months; start dates are flexible.
• Strong programming skills in Python, and experience with modern deep learning frameworks such as PyTorch or JAX.
• Hands-on experience designing, training, and debugging deep learning models for computer vision, 3D perception, or sequential decision-making.

Responsibilities

• Survey the latest research in scene understanding, prediction, and planning, and turn it into practical approaches that fit our system and constraints.
• Design and implement deep learning models that build a richer, real-time understanding of the driving environment from onboard sensors and other sources.
• Train and evaluate models on large-scale data from driving logs and public benchmarks, and lead failure analysis to identify where models fall short.
• Define metrics and evaluation methods that connect perception quality to actual vehicle behavior.
• Turn successful prototypes into solutions that work with our planning stack, with help from production engineers.
• Collaborate with Perception, Prediction, Planning, and Mapping teams so results move from experiment to product.

Skills

Python

Deep learning

Computer vision

Experiment design

Education

MS/PhD in CS/Robotics/EE/ML

Tools

PyTorch

JAX

Job description

About Didi Autonomous Driving

Founded in 2016, DiDi's autonomous driving unit is dedicated to developing Level 4 autonomous driving technology to make transportation safer, more efficient, and accessible. In 2019, the unit became an independent company, Voyager, to accelerate technology development, commercialization, and large-scale deployment. By combining DiDi's unique data advantage, operational expertise, and mobility ecosystem, we aim to build a reliable, scalable autonomous fleet that integrates seamlessly with human-driven services.

Driven by real-world customer needs and a strong commitment to safety and user experience, our mission is to deliver autonomous mobility solutions that create meaningful value for riders and communities.

About The Role

We want our vehicles to understand the physical world as it is right now, instead of relying on a map drawn long ago. Roads change, and the world a car actually drives in is always more complex than any static record of it. We are seeking a Research Intern to help give our planner a more accurate and up-to-date picture of the world, so the vehicle handles complex real-world scenes more robustly. In this role, you will take on hard, open problems from real-world driving and solve them with frontier research methods. You will report directly to the team lead and work on the problem from definition to validated solution. Strong results will often be worth publishing, and we will support you in writing them up, but the goal is to solve problems that matter to the product.

Responsibilities

• Survey the latest research in scene understanding, prediction, and planning, and turn it into practical approaches that fit our system and constraints.
• Design and implement deep learning models that build a richer, real-time understanding of the driving environment from onboard sensors and other available sources.
• Train and evaluate models on large-scale data from our driving logs and on public benchmarks, and lead failure analysis to find out where and why models fall short.
• Define metrics and evaluation methods that connect perception quality to how the vehicle actually drives.
• Turn successful prototypes into solutions that work with our planning stack, with help from production engineers.
• Collaborate with Perception, Prediction, Planning, and Mapping teams so results move from experiment to product.

Qualifications

• Currently enrolled in an M.S. or Ph.D. program in Computer Science, Robotics, Electrical Engineering, Machine Learning, or a related field, for the full duration of the internship.
• Available for a full-time onsite internship of 3–6 months; start dates are flexible.
• Strong programming skills in Python, and experience with modern deep learning frameworks such as PyTorch or JAX.
• Hands-on experience designing, training, and debugging deep learning models for computer vision, 3D perception, or sequential decision-making.
• Solid understanding of autonomous driving fundamentals, including perception, prediction, and motion planning.
• Ability to read research critically, form hypotheses, and turn them into well-controlled experiments.
• Excellent communication and collaboration skills.

Preferred Qualifications

• Publications at top venues such as CVPR, ICCV, ECCV, NeurIPS, ICLR, ICML, ICRA, IROS, CoRL or IV.
• Experience with multi-modal sensor fusion (camera, lidar, radar) or BEV representations.
• Experience with transformers, graph neural networks, or other structured scene representations.
• Experience with uncertainty estimation, or with learning-based prediction or planning.
• Experience with large-scale model training and data pipelines.
• Experience with public autonomous driving datasets such as nuScenes, Argoverse 2, or the Waymo Open Dataset.
• Proficiency in C++ (CUDA is a plus) and experience with robotics systems.

The hourly rate for the Intern position in the selected city is $46.

I acknowledge that prior to submitting this application, I have read and accepted the Privacy Notice for California Residents which is available on https://v.didi.cn/AQnxlBa