We believe there are many problems in the world in which robotics could play a significant role in making it easier, faster and safer for people to get things done.
You'd join a small scrappy technical team applying ML/RL to some hard unsolved problems combining robotics, controls, and sensing.
This is not a traditional team and involves substantial autonomy and ability to set your own technical direction, iterate, and prototype. Our team is a mix of long-term Googlers and people whose careers were primarily external software+hardware startups, and we have a minimal meetings/email/slides culture– we try to balance light planning and technical rigor with rolling up our sleeves and learning by experimenting and getting data.
Note: Practical engineering experience with reinforcement learning and related techniques is strongly desired. Experience with robotics (control systems, sensors, high-level applications) is also a benefit, but not required.
Note: This role involves work in an East Bay location (details TBD).
- Rapidly Learning, Building Prototypes, and Iterating
- Reading papers and talking with others across the ML (especially RL / Learning from Demonstrations / related) fields and evaluating which techniques, algorithms, architectures, or insights could be applicable tools to our problem space… then developing software/systems to implement and try them
- Running experiments (on physical robots, in simulators, or developing custom simplified simulators)
- Writing software (in Python, C++) and building on existing tools and libraries wherever that makes sense
- Analyzing data and using it to decide what to do next
- Figuring out how to reformulate / structure / parameterize problems to solve them (with a mix of first-principles engineering, optimization techniques, and learning techniques)
- Building and maintaining strong relationships with other engineers/researchers across Alphabet where their ML insights can move us forward or where we can provide valuable applications / data to create collaborations that benefit us both
- Being flexible and wearing several hats (e.g. helping identify and integrate new sensors when needed)
- Bachelor's degree in Computer Science or Robotics, or equivalent practical experience.
- Experience with machine learning algorithm development.
- Experience with one or more machine learning framework (e.g., TensorFlow, PyTorch, Caffe).
- Master's degree or PhD in Computer Science, Robotics or equivalent practical experience.
- Strong machine learning development experience in the field of robotics, simulation.
- Experience with using simulation for robotics or deep learning.
- Experience writing highly optimized and efficient code in Python, C, or C++.
About X, the moonshot factory
X creates radical new technologies to solve some of the world’s biggest problems. We develop uncomfortably ambitious, potentially world-changing new ideas such as self-driving cars, balloon-powered Internet and smart contact lenses. We’re a team of makers, entrepreneurs, engineers, designers and scientists with deep technical expertise who love the challenge of the seemingly impossible. X was formerly known as Google[x] and is part of Alphabet.