Responsible for automated and manual checks to ensure the vehicle software satisfies requirements. This includes code-level tests all the way to integration tests performed in simulated environments. Validation is a critical part of safe testing of our autonomous vehicle.

Key Responsibilities

• Manage the team's CI/CD infrastructure.

• Develop validation procedures to ensure the system behaves as expected.

• Work with teams to ensure all code is being tested properly.

Skills

• Experience with CI/CD pipelines.

• Strong software development abilities.

Responsible for creating tools that various sub-teams will use to create simulations and evaluate their results. These tools are critical to efficient software development across the team.

Key Responsibilities

• Work alongside many teams to understand their testing requirements.

• Develop and maintain tailored simulation solutions.

Skills

• Strong abilities to understand and implement feedback.

• Experience with user interface development.

Responsible for the software that finds a smooth trajectory for the vehicle that satisfies road rules, avoids obstacles, and is feasible to execute. As one of the last modules in the autonomous vehicle software, this code will ultimately determine where the vehicle moves.

Key Responsibilities

• Approximate the vehicle dynamics into simple models that can be evaluated efficiently.

• Design and implement algorithms that generate trajectory plans.

Skills

• Understanding of vehicle dynamics models and approximation methods.

• Knowledge of search algorithms.

• Strong software development abilities.

Responsible for the software that makes decisions about stop signs, traffic lights, and much more. They will work alongside the other leads to build a cohesive planning stack.

Key Responsibilities

• Build a complex and robust state machine.

• Conduct automated testing of various driving scenarios.

Skills

• Strong C++ programming skills.

• Ability to tackle highly complex engineering challenges.

• Experience with automated simulation tools.

Responsible for creating high fidelity 3D maps of environments based on driving data. This process involves

offline sensor fusion and state estimation such that large areas can be mapped accurately.

Key Responsibilities

• Building a mapping pipeline which produces 3D maps from sensor data.

• Removing dynamic objects from maps.

• Working alongside the semantic mapping team to provide sufficient information for labelling.

Skills

• Understanding of state estimation algorithms.

• Experience working with large datasets.

Responsible for augmenting the 3D maps with labels such as lane lines, curbs, traffic lights, etc. This involves creating a streamlined process for turning raw 3D data into information-rich maps that can be utilized across the autonomy software.

Key Responsibilities

• Creating an interface for labeling semantic information on 3D maps.

• Working closely with subteams to understand their requirements.

• Maintaining clear processes and guidelines to avoid errors and ambiguities in maps.

Skills

• Experience with user interface development.

• Ability to understand and communicate complex technical requirements.

Responsible for building models that run directly on camera, LiDAR, and RADAR data to produce detections. These models typically exhibit higher performance at the expense of added complexity.

Key Responsibilities

• Researching SOTA detection models with an emphasis on real-time performance.

• Train models and deploy experiments to evaluate the performance compared to existing 2DOD and 3DOD models.

Skills

• Strong understanding of deep neural networks.

• Experience optimizing model performance.

Responsible for the software that processes multimodal detections into smoothed outputs. Tracking is a key step in the robotics pipeline, bridging the gap between raw detections and object information suitable for trajectory planning.

Key Responsibilities

• Create a robust tracker that can operate at high speeds.

• Understand and account for the nuances of various detection methods.

• Work alongside the planning team to satisfy object detection requirements.

Skills

• Understanding of state estimation algorithms.

• Experience with LiDAR, camera, and/or RADAR object detection.

• Strong C++ programming skills.

Responsible for creating models that predict the future behaviour of road users based on their history. Accurate prediction allows the vehicle to navigate safely in complex environments.

Key Responsibilities

• Explore and benchmark motion prediction algorithms.

• Understand and meet requirements such as latency and accuracy.

• Train and evaluate motion prediction models.

Skills

• Experience training and evaluating machine learning models.

• Strong understanding of deep neural networks.

• Experience optimizing model performance.

Responsible for developing and maintaining the infrastructure for data storage, data labelling, and model training.

Key Responsibilities

• Create and manage the infrastructure to house our internal datasets and train models.

• Work with perception teams to accommodate their data needs.

Skills

• Experience with large datasets and/or shared compute infrastructure.

• Understanding of data storage and compression concepts such as video encoding.

Responsible for optimizing machine learning models for the best performance on our vehicle. Both low latency and high throughput are crucial to safe autonomous driving.

Key Responsibilities

• Evaluate the performance of machine learning models.

• Work alongside perception teams to optimize model architecture.

• Compile models and manage the data pipeline to optimize performance.

Skills

• Experience with acceleration libraries such as TensorRT.

• Understanding of inter-process communication methods for large amounts of data.