PhD Residency - AI / ML : Language Model Interpretability

Key Skills

Java Programming, Python Programming, C++, C Programming, Keras software library, TensorFlow

Job Description

X is Alphabet’s moonshot factory. We are a diverse group of inventors and entrepreneurs who build and launch technologies that aim to improve the lives of millions, even billions, of people. Our goal: 10x impact on the world’s most intractable problems, not just 10% improvement. We approach projects that have the aspiration and riskiness of research with the speed and ambition of a startup.

THE FOUNDATION OF AN AMAZING JOURNEY

Our goal at X is to make the world a radically better place. In order to do that we seek fresh unexpected perspectives, from different fields, and that’s why we’re excited about you.

Life here isn’t easy, but it’s fun. We’re trying to build things most people can’t even imagine, and we’re doing it with the hope of making a huge, positive impact on the world. You’ll be embedded into a moonshot project, where you’ll partner with team members to solve key challenges.

This isn’t your ordinary internship. You’ll be positively challenged and pushed professionally, in ways that you may have never experienced. If this excites you - keep reading.

DURING YOUR INTERNSHIP YOU CAN EXPECT:

• To be placed on one of our confidential or public X projects
• To get paid competitively and with Google benefits
• To be part of a lively community of other Interns and Residents
• To addend colloquium and discussions with team leads from across Google, DeepMind and external organizations

DETAILS:

• Location: X's headquarters in Mountain View, CA
• Start Date(s): Year-round rolling basis
• Duration: a flexible 4 mo. to 1 year program based on project team needs and your availability

REQUIREMENTS:

• Must be enrolled in an academic program and working towards completing a PhD degree

WHAT YOU'LL BE WORKING ON:

• Apply existing methods for language model interpretation (integrated gradients, attention visualization, etc.) to language models trained on code.
• Develop new interpretability methods to take advantage of code structure.
• Develop visualization tools and interfaces to enable researchers and users to interact with the tools you build.

MINIMUM QUALIFICATIONS:

• Currently enrolled in a PhD program in Computer Science with a strong interest in machine learning, interpretability and software productivity
• Experience implementing interpretability methods for neural models
• A fast learner with strong coding skills in Java/Python or equivalent
• Passionate about development and the future of software engineering
• Ability to work in a small team / rapid prototyping environment and deal with uncertainty and fluidity

PREFERRED QUALIFICATIONS:

• Open-source projects that demonstrate relevant skills and/or publications in relevant conferences and journals (NeurIPS, ICLR or ICML, OOPSLA, PLDI, POPL, ICSE/FSE)
• Aptitude in TensorFlow and/or Jax
• Experience in data visualization and/or user interface design.
• Familiarity with programming language theory/practices and/or static code analysis
• Experience working in a startup/research environment