Research Scientist, Verified Code Generation, DeepMind

Research Scientist, Verified Code Generation, DeepMind Overview

This role sits within a research-focused AI organization that combines fundamental research, product innovation, and infrastructure work. The team gives researchers room to concentrate on different kinds of work depending on the project, while expecting them to move ideas into practice quickly, run large-scale experiments, and manage deliverables. The broader mission is to apply advanced research to real-world problems across computer science, contribute to new products and technologies, and share results with the wider research community through publication and collaboration.

The position is centered on verified code generation and formal methods, with a strong emphasis on using the Lean proof assistant to reason about correctness, safety, and security. The work also connects to large language models, code analysis, compiler infrastructure, and systems security. The organization describes itself as an interdisciplinary AI lab focused on advancing AI in ways that support scientific discovery, improve products for billions of users, and keep safety and ethics at the forefront.

What you will do

• Build AI systems that can produce formal proofs demonstrating that software is correct, safe, and secure, using Lean as the main proof environment.
• Express programming language semantics in Lean so that verified static analysis can be applied to substantial, real-world codebases.
• Design, prototype, and assess new methods that combine large language models with formal verification to support automated code analysis and generation.
• Create tools, libraries, and infrastructure that allow formal verification workflows to scale to large codebases and production settings.
• Partner with researchers and engineers across AI, security, and compiler infrastructure teams to develop and apply these techniques.
• Run large-scale tests, move promising ideas into deployment rapidly, and keep work aligned with deadlines and deliverables.
• Contribute to the research community by publishing results and sharing insights developed through internal work and external academic partnerships.

What the role requires

• A PhD in computer science, programming languages, formal methods, software engineering, or a closely related technical field, or equivalent practical experience.
• Four or more years of experience in at least one of these areas: programming language semantics, static analysis, abstract interpretation, software verification, or interactive theorem proving.
• At least one year of hands-on experience with a proof assistant such as Lean, Coq, Isabelle, or a similar tool.
• Two or more years of experience with compiler infrastructure such as LLVM or GCC, or with formalizing programming languages.
• At least one year of experience applying large language models or machine learning to code or reasoning tasks.
• At least one year of experience in memory-safety analysis, vulnerability research, or systems security.
• At least one year of experience working with Lean 4.
• Proven experience building and scaling software verification tools for production codebases.
• A publication history in leading formal methods or software security conferences or journals such as POPL, PLDI, CCS, or S&P.
• English proficiency, since the company states that English is required for all roles unless the posting explicitly says otherwise.

Compensation and perks

• For the US location, the posted base salary range is $207,000 to $301,000.
• The package also includes a 20% bonus target.
• Equity is included.
• Benefits are provided, and the posting directs candidates to learn more about Google benefits.
• Applicants can indicate a preferred location from London, UK or Mountain View, CA, USA.
• The role offers the chance to work in a global research environment with strong interdisciplinary collaboration and opportunities for learning and career growth.

Additional information

The employer states that it is an equal opportunity and affirmative action employer and that it is committed to building a workforce that reflects the users it serves. It also notes that reasonable accommodations are available for applicants who need them.

The posting says that Google does not accept agency resumes and asks agencies not to send resumes to job aliases, employees, or other locations. It also explains that unsolicited resumes do not create any fee obligation for Google.

The company further notes that equity grants are handled by Alphabet Inc. under the terms of the relevant stock plan and grant agreement, are discretionary, and are not guaranteed.

Applicants should be aware that the posting says English proficiency is required for all roles unless the job ad states otherwise.

Degree Requirement: Bachelor’s

Visa Sponsorship May be