Senior Research Engineer

Aisafety; Berkeley (CA); On-site; USD 150,000 - 250,000; Full time

Benefits offered by this job

Catered lunch and dinner

Visa sponsorship for in-person employees

Work-related travel expenses covered

Job summary

An AI safety research organization in Berkeley is looking for a Senior Research Engineer to develop and lead projects aimed at enhancing AI safety. This role requires significant software engineering experience, fluency in Python, and expertise in areas such as machine learning, high-performance computing, or technical leadership. Both remote and in-person work arrangements are available. The position offers a competitive salary range of $150,000-$250,000/year based on experience.

Qualifications

• Significant software engineering experience, preferably with open-source contributions.
• Fluency in Python is a must.
• Must be results-oriented and motivated by impactful research.

Responsibilities

• Detect deceptive behavior in AI models and develop mitigation strategies.
• Lead projects on preventing catastrophic misuse of AI.
• Build frameworks for research to accelerate experiments.

Skills

Python proficiency

Experience in training transformers

High-performance computing

Project management

Technical leadership

Knowledge of linear algebra and statistics

Tools

PyTorch

Kubernetes

Job description

About Us

FAR.AI is a non-profit AI research institute dedicated to ensuring advanced AI is safe and beneficial for everyone. Our mission is to facilitate breakthrough AI safety research, advance global understanding of AI risks and solutions, and foster a coordinated global response. Founded in July 2022, we have grown quickly to 30+ staff. We are uniquely positioned to conduct technical research at a scale surpassing academia and leveraging the research freedom of being a non-profit. Our work is published at top conferences (e.g. NeurIPS, ICLR, ICML) and cited by leading media outlets such as the Financial Times, Nature News and MIT Technology Review.

FAR.AI uses three prongs working together to improve AI safety:

• FAR.Research - we conduct cutting-edge AI safety research in-house and dispense grants to support the wider research community.
• FAR.Futures - we bring together key policy makers, researchers and companies to drive change, such as the San Diego Alignment Workshop or the Guaranteed Safe AI research roadmap written with Yoshua Bengio.
• FAR.Labs - we host a co-working space in Berkeley to help incubate other AI safety organizations, currently housing 40 members.

About FAR.Research

We explore promising research directions in AI safety and scale up only those showing a high potential for impact. Once the core research problems are solved, we work to scale them to a minimum viable prototype, demonstrating their validity to AI companies and governments to drive adoption.

We are aiming to rapidly grow our team in the following areas:

• Mitigating AI deception: Studying when lie detectors induce honesty or evasion, and developing for deception and sandbagging
• Evals and red-teaming: Conducting pre- and post-release adversarial evaluations of frontier models (e.g. Claude 4 Opus, ChatGPT Agent, GPT-5); developing novel attacks to support this work; and exploring new threat models (e.g. persuasion, tampering risks).
• Infrastructure: Maintaining GPU compute infrastructure to support experiments with open-weight models and developing new tooling to allow our research teams to scale their fine-tuning and post-training workflows to frontier open-weight models.
• Adversarial Robustness: Working to rigorously solve these security problems through building a science of security and robustness for AI, from demonstrating superhuman systems can be vulnerable, to scaling laws for robustness and jailbreaking constitutional classifiers.
• Mechanistic Interpretability: Finding issues with Sparse Autoencoders, probing deception using AmongUs, understanding learned planning in SokoBan and interpretable data attribution.

FAR.AI is one of the largest independent AI safety research institutes, and is rapidly growing with the goal of diversifying and deepening our research portfolio. We would welcome the opportunity to add new research directions if you are a senior researcher with a strong vision and would like to pitch us on it.

About the Role

This role would be a good fit for an experienced machine learning engineer, or an experienced software engineer looking to transition to AI safety research. All candidates are expected to:

• Have significant software engineering experience. Evidence of this may include prior work experience and open-source contributions.
• Be fluent working in Python.
• Be results-oriented and motivated by impactful research.
• Bring prior experience mentoring other engineers or scientists in engineering skills.

Additionally, candidates are expected to bring expertise in one of the following areas corresponding to the core competencies our different research teams most need:

• Option 1 – Machine Learning:
• Substantial experience training transformers with common ML frameworks like PyTorch or Jax.
• Good knowledge of basic linear algebra, calculus, vector probability, and statistics.
• Option 2 – High-Performance Computing:
• Power user of cluster orchestrators such as Kubernetes (preferred) or SLURM
• Experience building high-performance distributed-systems (e.g. multi-node training, large-scale numerical computation)
• Experience optimizing and profiling code (ideally including on GPU, e.g. CUDA kernels).
• Option 3 – Technical Leadership:
• Experience designing large-scale software systems, whether as an architect in greenfield software development or leading a major refactor.
• Comfortable project managing small teams, such as chairing stand-ups and developing detailed roadmaps to execute on a 3-6 month research vision.

About the Projects

As a Member of Technical Staff (Senior Research Engineer) you would join one of our existing workstreams and lead projects there:

• Detecting and preventing deception. Under what conditions can we reliably detect deceptive behaviour from models, and can such behaviour be effectively mitigated at scale? This would focus on large-scale training of transformers.
• Preventing catastrophic misuse. Apply our research insights to detect and mitigate vulnerabilities and other risks in frontier AI models. This would focus more on technical leadership
• Accelerating our research. Build frameworks and infrastructure that allows us to ask bigger questions and more rapidly run new experiments, to deepen our research. This would focus more on high-performance computing.

As we continue to grow our research portfolio, additional workstreams may open up for contribution, for example in mechanistic interpretability.

Logistics

If based in the USA, you will be an employee of FAR.AI, a 501(c)(3) research non-profit. Outside the USA, you will be an employee of an EoR organization on behalf of FAR.AI.

• Location: Both remote (global) and in-person (Berkeley, CA) are possible. We sponsor visas for in-person employees, and can also hire remotely in most countries and time zones, provided you are willing to overlap for 2h of the Berkeley working day.

• Hours: Full-time (40 hours/week).

• Compensation: $150,000-$250,000/year depending on experience and location, with the potential for additional compensation for exceptional candidates. We will also pay for work-related travel and equipment expenses. We offer catered lunch and dinner at our offices in Berkeley.

• Application process: A 72-minute programming assessment, two interviews with members of our technical staff, and a 1-2 week paid work trial. If you are not available for a work trial we may be able to find alternative ways of testing your fit.

If you have any questions about the role, please do get in touch at talent at far.ai.

Otherwise, if you don\'t have questions, the best way to ensure a proper review of your skills and qualifications is by applying directly via the application form. Please don\'t email us to share your resume (it won\'t have any impact on our decision). Thank you!