Machine Learning Research Engineer

Job description

We're building the first truly private, personal AI that learns your skills, judgment, and preferences without big tech—or us—ever seeing your data. Our core ML challenge: how do we train the world’s best personal models? WHAT YOU'LL DO • Train models to think, act, and value like a specific user. Implement & test parameter-efficient fine-tuning techniques to make models pick up people’s reasoning & judgement & style. • Put small data to use. Create synthetic data pipelines to let models squeeze everything out of small datasets. Invent new data collection pipelines (working with the product team), test their effects on model quality, and make recommendations on what sort of data matters most to collect. • Think carefully about how to evaluate quality. We're not trying to push up numbers on a public benchmark, we’re trying to make models qualitatively good at understanding a particular user’s taste & judgement. And we’re not just finetuning one model, but building a system that can finetune a quality private model for anyone. • Reduce misuse and catastrophic risk while maintaining bulletproof privacy. Develop guardrails to prevent models from gaining capabilities that are illegal or pose society-wide dangers. YOU HAVE • A deep understanding of transformers and parameter-efficient finetuning. You can dive into the details of how transformers work, the tradeoffs and considerations in using parameter-efficient finetuning techniques, and you have good intuitions for how datasets affect model behaviour. • Ability to execute quickly. We ship fast and fail fast so we can win faster. The challenge of human relevance in a post-AGI world isn’t going to solve itself. • A missionary mentality. We’re a mission-driven company, looking for mission-first people. If you’re passionate about ensuring AI works for people (and not the other way around), you’ve come to the right place. • Ready to roll up your sleeves. We're an early stage startup, so we’re looking for someone who can wear many hats. You’ll put ML theory into practice. EXPERIENCE YOU MAY HAVE • Experience running finetuning or post-training experiments with LLMs. • Work at a fast-paced AI startup, or top AI lab. • Published machine learning research outputs, as peer-reviewed papers or in-depth technical blog posts. • Academic experience in machine learning research. This might be through a PhD, a research-oriented master’s, or research programs like MATS or Anthropic Fellows Program. We encourage speculative applications; we expect many strong candidates will have different experience or unconventional backgrounds. WHAT WE OFFER • Generous compensation and early stage equity. We’re competitive with the top startups, because we believe the best talent deserves it. • World-class expertise. We’re based in top AI research hubs in San Francisco and London. We're backed by AI experts like Juniper Ventures, Seldon Lab, and angels at Anthropic and Apollo Research. You’ll have access to some of the best AI expertise in the world. • Massive impact. Our mission is to keep people in the economy well after AGI. You’ll help shift the trajectory of AI development for the better, helping break the intelligence curse and prevent gradual disempowerment to keep humans in control of the future. ABOUT WORKSHOP LABS We’re building the AI economy for humans. While everyone else tries to automate the world top-down, we believe in augmenting people bottom-up. Our team previously created evals used by Open AI, completed frontier AI research at MIT/Cambridge/Oxford, worked in Stuart Russell's lab, and led product verticals at high growth startups. Our founders' essay series on AI’s economic impacts and the technology we’ll need to keep humans relevant, The Intelligence Curse https://intelligence-curse.ai/, has been covered in TIME https://time.com/7289692/when-ai-replaces-workers/, The New York Times https://www.nytimes.com/2025/05/15/opinion/artifical-intelligence-2027.html, and AI 2027 https://ai-2027.com/footnotes#footnote-slowdown-51. Our vision is for everyone to have a personal AI aligned to their goals and values, helping them stay durably relevant in a post-AGI economy. As a public benefit corporation, we have a fiduciary duty to ensure that as AI becomes more powerful, humans become more empowered, not disempowered or replaced. We’re an early stage startup, backed by legendary investors like Brad Burnham and Matt McIlwain, visionary product leaders like Jake Knapp and John Zeratsky, philosopher-builders like Brendan McCord, and top AI safety funds like Juniper Ventures. Our investors were early at Anthropic, Slack, Prime Intellect, DuckDuckGo, and Goodfire. Our advisors have held senior roles at Anthropic, Google DeepMind, and UK AISI.