Research Engineer

About the role

FAR.AI organizes Members of Technical Staff to conduct safety research on advanced AI and to build implementations that run experiments. In this Research Engineer role, you will own the design and execution of machine learning algorithms that directly power scientific experiments aimed at improving AI safety. You will work hand in hand with research scientists to translate theoretical safety concepts into functional code and measurable laboratory outcomes. The position requires you to take ownership of the full experiment lifecycle, from hypothesis formulation through implementation, debugging, and rigorous analysis of results. You will be responsible for developing the engineering infrastructure that allows sparse autoencoders and learned planners to be tested at scale. A core part of your work will involve running adversarial evaluations of frontier models such as Claude 4 Opus, ChatGPT Agent, and GPT-5 to probe model behaviors. You will translate research directions into concrete technical proposals that can be pitched to diversify and deepen the institute portfolio. Finally, you will collaborate closely with scientists to produce findings destined for premier academic venues, ensuring your engineering work is recognized in published papers.

Key facts

What you'll do

Implement machine learning algorithms that execute experiments to advance AI safety research and validate theoretical models.

Run adversarial evaluations and red-teaming exercises against frontier models including Claude 4 Opus, ChatGPT Agent, and GPT-5 to measure safety-relevant capabilities and failures.

Investigate mechanistic interpretability challenges by building and testing sparse autoencoders designed to uncover latent representations.

Develop learned planners that enable agents to reason over complex safety constraints and long-horizon tasks within controlled environments.

Propose novel research directions and formally pitch these ideas to expand the institute portfolio and align with broader AI safety priorities.

Write clear scientific narratives and collaborate with researchers to produce papers suitable for premier venues in AI and machine learning.

Maintain scalable machine learning infrastructure that supports large-scale experimentation, data logging, and reproducibility.

Contribute to open source libraries where appropriate to ensure that implemented techniques benefit the wider research community.

Participate in internal code reviews and pair programming sessions to elevate code quality and ensure robust experiment tracking.

Document every phase of the research process so that experiments can be audited, replicated, and understood by peers.

Requirements

Candidates must hold a degree that meets the standard stated in the source metadata for this position.

You must have hands-on experience with GPU compute infrastructure and large-scale fine-tuning workflows for modern neural networks.

Demonstrate strong engineering skills that allow you to build reliable, scalable machine learning implementations under tight deadlines.

You must have a track record of contributing to open source projects relevant to machine learning and scientific computing.

The ability to author scientific papers and collaborate closely with researchers is required to earn credited authorship on publications.

You should be comfortable working with transformer libraries and PyTorch based stacks that power current AI safety experiments.

Experience with red-teaming, threat modeling, and adversarial evaluation of powerful language models is essential.

You must be able to work within the Berkeley co-working environment and adhere to the standards of the Members of Technical Staff team.

Nice to have

Only items explicitly indicated as preferred in the source material are listed here.

Practical notes

This position is based in the Berkeley Office and requires active participation as a Member of Technical Staff.

Typical interview steps

Data interviews commonly include a SQL or coding exercise, a statistics question, and a case study. Candidates may be asked to design a metric, interpret an experiment, or build a small model. Some companies give a take-home analysis. Expect questions about past projects and the business impact of your work. Interviewers often evaluate how you communicate uncertainty and business impact, not only the math. Bringing a clean write-up of a past analysis to the interview is well received.

Good to know

Research engineers bridge scientific investigation and production implementation in AI safety.

Machine learning frameworks such as PyTorch and transformer libraries are central to daily work.

Collaboration with academic conferences and government evaluations is common in AI safety research.

Infrastructure tooling supports large scale experimentation with open weight models.

The position involves both independent research and structured mentorship through code review and pair programming.

Questions to ask

Worth asking in any interview: how the team measures success, who the role works with daily, what the onboarding looks like, and what the company is trying to achieve this year. Asking what past hires did well is a strong final question. Keep the list short and pick the questions that matter most to you.

Career growth

Data careers grow toward senior analyst, staff data scientist, or data engineering lead. Many professionals specialize in machine learning, analytics, or infrastructure. Cross-functional work with product and engineering teams becomes more important at senior levels. The field changes quickly, so continuous learning is part of the job. Professionals who can translate numbers into decisions tend to advance fastest.

About the company

FAR. AI http://FAR. AI is a non-profit AI research institute working to ensure advanced AI is safe and beneficial for everyone.