The Traffic Lights and X-Object understanding team works to build robust capable models that help the Cruise vehicles understand and safely interact with the world around it. These models help classify fine grained semantic information concerning Vulnerable Road Users (VRUs) and other critical objects in the vehicle's environment, identify and comply with directives from humans controlling traffic or from Traffic control devices, such as Traffic lights or caution tape. The team helps ensure a high safety bar as well as predictable interactions with VRUs and comfortable driving, crucial for both legal compliance and demonstrating responsible roadmanship.
We are looking for a Machine Learning Engineer in Perception that will invent pragmatic, long-term and stable software solutions to sophisticated problems. This adaptable role implements a wide spectrum of applied ML and sensor research, and model development to overcome challenges on urban roads.
Ideal candidates would possess a strong background in machine learning, particularly in developing algorithms for perception and overall machine learning workflow. Experience in multi-task learning, feature extraction, and model architecture as well as working with downstream models to leverage perception signals is highly valued. The team seeks individuals who are passionate about improving road safety for VRUs and are eager to contribute to the development of innovative solutions that enhance Cruise's autonomous vehicle technology.
What you’ll be doing:
- From a research perspective, exploring, prototyping, validating, and iterating new algorithms
- On the development side, driving efforts to optimize, supervise and refine on-road performance for models on the path to production
- This role will support the hardware organization. Likely projects will include sensitivity studies for hardware selection as well as architecture investigations.
- Champion of the highest quality in engineering, crafting solutions and not just identifying problems
What you must have:
- Solid understanding of computer vision fundamentals including the development and training of ML models
- Perform first principles analysis of perception models and experiments to support HW sensor development.
- Be able to run and retrain Cruise perception machine learning models and evaluate model performance (PR, other metrics)
- Must be able to work independently and multi-functionally between HW and perception teams.
- Willing to be hands on with sensor data; has a practical understanding of experimental data and hardware versus purely theoretical.
- Basic knowledge of sensing physics and awareness of different sensing modalities (lidar, cameras, radars, ultrasonic sensors, etc.).
- Growing expertise with ML frameworks such as Tensorflow, Caffe, and PyTorch
- Strong programming skills in Python, C++, or other relevant languages
- Passionate about self-driving car technology and its impact on the world!
Bonus points!
- PhD in machine learning or computer science
- Research experience developing computational cameras, physics-based computer vision systems, or other research at the intersection of CV/ML and sensor development.
- 4+ years of industry experience in Machine Learning
- Experience and growing expertise with ROS, OpenCV, Gazebo, or PCL
- Track record in deploying perception algorithms into real world environments
- Experience and growing expertise with autonomous driving technologies
The salary range for this position is $161,200 - $237,000. Compensation will vary depending on location, job-related knowledge, skills, and experience. You may also be offered a bonus, long-term incentives, and benefits. These ranges are subject to change.
