Research Engineer / Scientist, Anthropic
Alignment Science
£230-515k
+ Equity
More information about location
AI safety and research company
Open for applications
AI safety and research company
1001+ employees
Open for applications
£230-515k
+ Equity
More information about location
1001+ employees
Company mission
To create reliable, interpretable, and steerable AI systems.
Role
Who you are
- Have significant software, ML, or research engineering experience
- Have some experience contributing to empirical AI research projects
- Have some familiarity with technical AI safety research
- Prefer fast-moving collaborative projects to extensive solo efforts
- Pick up slack, even if it goes outside your job description
- Care about the impacts of AI
Desirable
- Have experience authoring research papers in machine learning, NLP, or AI safety
- Have experience with LLMs
- Have experience with reinforcement learning
- Have experience with Kubernetes clusters and complex shared codebases
What the job involves
- As a Research Engineer on Alignment Science, you'll contribute to exploratory experimental research on AI safety, with a focus on risks from powerful future systems (like those we would designate as ASL-3 or ASL-4 under our Responsible Scaling Policy), often in collaboration with other teams including Interpretability, Fine-Tuning, and the Frontier Red Team
- Testing the robustness of our safety techniques by training language models to subvert our safety techniques, and seeing how effective they are at subverting our interventions
- Run multi-agent reinforcement learning experiments to test out techniques like AI Debate
- Build tooling to efficiently evaluate the effectiveness of novel LLM-generated jailbreaks
- Write scripts and prompts to efficiently produce evaluation questions to test models’ reasoning abilities in safety-relevant contexts
- Contribute ideas, figures, and writing to research papers, blog posts, and talks
- Run experiments that feed into key AI safety efforts at Anthropic, like the design and implementation of our Responsible Scaling Policy
Salary benchmarks
Our take
Funded by Google and founded by ex-OpenAI employees, Anthropic is an AI company focused on building a reliable, innovative AI system. While the company is relatively new to the market it has already launched a highly anticipated AI chat assistant called Claude which is set to rival OpenAI's ChatGPT.
Similar to other chatbots, Claude is accessible through a chat interface and can provide a number of conversation and text-processing tasks. A key point of differentiation is that Claude is designed to produce less harmful outputs than that of the other chatbots that came before it.
While there is much praise for the chat assistant, there is still room for improvement with feedback reporting that Claude is worse at maths and grammar compared to its rival. However, with industry giants like Google backing this company from an early stage, it is clear Anthropic has the potential to become a leading player within the AI industry.
Freddie
Company Specialist at Welcome to the Jungle
Insights
Few candidates hear
back within 2 weeks
190% employee growth in 12 months
Glassdoor (4.3)
Trustpilot (2)
Company
Funding (last 2 of 11 rounds)
Jan 2025
$1bn
GROWTH EQUITY VC
Nov 2024
$4bn
GROWTH EQUITY VC
Total funding: $12.8bn
Company benefits
- Comprehensive health, dental, and vision insurance for you and your dependents
- Inclusive fertility benefits via Carrot Fertility
- Generous subsidy for OneMedical
- 21 weeks of paid parental leave
- Unlimited PTO
- Optional equity donation matching at a 3:1 ratio, up to 50% of your equity grant
- 401(k) plan with 4% matching
- $500/month flexible wellness stipend
- Commuter coverage
- Annual education stipend
- A home office improvement stipend when you first join
- Relocation support for those moving to the Bay Area
Company values
- Here for the mission: We pursue our mission by building frontier systems, studying their behaviors, working to responsibly deploy them, and regularly sharing our safety insights. We collaborate with other projects and stakeholders seeking a similar outcome.
- Unusually high trust: We assume good faith, disagree kindly, and prioritize honesty. We expect emotional maturity and intellectual openness. At its best, our trust enables us to make better decisions as an organization than any one of us could as individuals.
- Do the simple thing that works: We celebrate trying the simple thing before the clever, novel thing. We embrace pragmatism - sensible, practical approaches that acknowledge tradeoffs. We love empiricism - finding out what actually works by trying it - and apply this to our research, our engineering and our collaboration. We aim to be open about what we understand and what we don’t.
- One big team: Collaboration is central to our work, culture, and value proposition. While we have many teams at Anthropic, we feel the broader sense in which we are all on the same team working together towards the mission.
Company HQ
SoMa, San Francisco, CA
Articles
Leadership
Dario Amodei
(CEO & Co-Founder)Spent almost 5 years at OpenAI as a Team Lead for AI safety and as the Vice President of Research. Also worked as a Research Scientist at Google.
Jared Kaplan
(Chief Science Officer & Co-Founder)Has been a Professor at John Hopkins University for the past 10 years. Also an experienced researcher having worked previously at OpenAI and the SLAC National Accelerator Laboratory.
Jack Clark
(Co-Founder)Experienced reporter having worked previously at the Register and Bloomberg. Also worked for OpenAI in Strategy and Communications and as a Policy Director.
Sam McCandlish
(Co-Founder)Graduated from Stanford University with a PhD in Theoretical Physics. Previously worked as a Research Lead at OpenAI.
Tom Brown
(Co-Founder)Previously worked as a Member of the Technical Staff for both OpenAI and Google Brain.
Share this job
View 113 more jobs at Anthropic