Research Engineer Visa Sponsorship Uk

Job Description

JOB DESCRIPTION

You want to help build and quickly iterate on machine learning experiments to help us improve the behaviour of powerful AI systems through finetuning. You are concerned with making AI more helpful, honest, and harmless, as well as with influencing model behaviour to better line with human values and aims. You may identify as both a scientist and an engineer. As a Research Scientist or Research Engineer on the Finetuning team, you will help to improve language models using techniques such as constitutional AI. You will be able to conduct innovative research on frontier models and see your efforts translate into tangible gains in performance and safety.

We want researchers to be able to iterate on their own investigations. We also offer possibilities for engineers to conduct their own research initiatives. As a result, depending on the candidate’s skills and interests, this post can be more research or engineering focused.

Note: Currently, the team prefers individuals who can be based in the Bay Area. However, we are still open to any candidate who can meet the organization’s 25% in-person requirement.

About Anthropic

Anthropic is an AI safety and research business dedicated to developing trustworthy, interpretable, and steerable AI systems. We want AI to be safe and helpful for both our consumers and society as a whole. Our interdisciplinary team has expertise in machine learning, physics, philosophy, and computer science.

Representative projects:

• Help develop unique fine-tuning strategies to improve language model behaviour and make models more useful, honest, and harmless.

• Test out strategies like constitutional AI at scale and measure their effects on model behaviour.

• Create tools and infrastructure to facilitate efficient fine-tuning experiments on big language models.

• Create unique prompts and ways for improving and testing model behaviours.

• Conduct experiments that contribute to vital AI research and safety projects at Anthropic.

You could be a good fit if:

• Possess extensive Python, machine learning, research engineering, or research experience.
• Prefer fast-paced collaborative projects with defined goals, such as improving model behaviours.
• Are results-oriented, with a propensity towards flexibility and impact.
• Pick up slack, even if it is outside of your job description.
• Concerned about the influence of AI and your work.

A strong candidate may also:

• Have previous familiarity with big language model finetuning approaches such as RLHF.
• Have experience with complicated shared code bases and RL infrastructure.
• Have experience writing academic papers in machine learning, natural language processing, or AI alignment, or related industry experience.

Annual salary (USD)

• The projected salary range for this position is $280,000 to $600,000 USD.

Logistics

Our location-based hybrid policy requires personnel to spend at least 25% of their time in the office.

Deadline for applying: None. Applications will be considered on a rolling basis.

We offer US visa sponsorship. However, we are unable to effectively sponsor visas for every post and candidate; operations positions are particularly challenging to support. But, if we make you an offer, we will make every attempt to get you into the United States, and we have hired an immigration lawyer to assist us.

We encourage you to apply even if you don’t think you meet every single need.Not all strong candidates will meet all of the qualifications given. According to research, persons who identify as members of underrepresented groups are more likely to experience imposter syndrome and doubt the validity of their candidature, thus we encourage you not to rule yourself out early and to apply if you are interested in this work. We believe that AI systems like the ones we’re developing have huge social and ethical consequences. We believe that this emphasises the importance of representation, and we endeavour to have a diverse team.

Compensation and Benefits*

Anthropic’s remuneration structure includes three components: salary, stock, and benefits. We are committed to paying fairly and intend for these three factors to be extremely competitive with market prices.

Equity will make up a significant portion of overall remuneration for this position, in addition to the income indicated above. We intend to provide higher-than-average stock remuneration for a company of our size and will disclose equity levels at the time of offer issuance.

Our US-based workers receive the following benefits:

• Optional equity donation matching at a 3:1 ratio, for up to 50% of your equity grant.
• Comprehensive health, dental, and vision coverage for you and your dependents.
• 401(k) plan with 4% match.
• 22 weeks of compensated parental leave.
• Unlimited PTO – most employees take 4-6 weeks per year, sometimes more!
• Stipends for schooling, home office renovations, transportation, and wellbeing.
• Carrot has fertility advantages.
• Our office serves daily lunches and snacks.
• Relocation assistance for folks migrating to the Bay Area.

UK Benefits: The following benefits are available to our UK-based employees:

• Optional equity donation matching at a 3:1 ratio, for up to 50% of your equity grant.
• Private health, dental, and vision insurance for yourself and your dependents.
• Pension contribution (4% of your earnings).
• 22 weeks of compensated parental leave.
• Unlimited PTO – most employees take 4-6 weeks per year, sometimes more!
• Health cash plan.
• Life insurance and income protection.
• Our office serves daily lunches and snacks.

This pay and benefits information is based on Anthropic’s best estimate for this position as of the date of publishing and may be updated in the future. Employees based in countries other than the United Kingdom or the United States will receive a different benefits package. The amount of remuneration within the range will be determined by a number of job-related elements, including your position on our internal performance ladders, which are based on factors such as previous work experience, applicable education, and performance during our interviews or work trials.

How we are different.

We believe that the most influential AI research will be big science. At Anthropic, we work as a coherent team on a few large-scale research projects. And we prioritise effect – achieving our long-term goals of steerable, trustworthy AI — over working on smaller, more particular issues. We see AI research as an empirical discipline that shares many similarities with physics and biology, as well as traditional computer science activities. We are a very collaborative group, and we hold frequent research discussions to ensure that we are working on the most impactful projects at all times. As such, we place a high priority on communication abilities. We don’t distinguish between engineering and research, and we want all of our technical team to contribute to both as needed.

The simplest approach to comprehend our research directions is to read our most recent findings. This study builds on many of the directions our team worked on prior to Anthropic, such as GPT-3, Circuit-Based Interpretability, Multimodal Neurons, Scaling Laws, AI & Compute, Concrete Problems in AI Safety, and Learning from Human Preferences.

Come to work with us!

Anthropic is a public benefit corporation headquartered in San Francisco. We provide competitive remuneration and benefits, optional equity donation matching, substantial vacation and parental leave, flexible working hours, and a great office space for collaboration with colleagues.