FUTURESEARCH is looking for exceptional Research Scientists to figure out the best way to research and reason about hard, judgment-laden questions.
We are an elite team of engineers and researchers confronting the frontier of AI-based reasoning, with deep backgrounds in the art and science of forecasting.
Examples of our recent work include Deep Research Bench, a benchmark for evaluating LLM performance on Deep Research tasks, and Bench to the Future, a pastcasting benchmark that is testing the ability of agents to make predictions.
Requirements
You are an experienced researcher in math, physics, CS, or an equivalently quantitative field. You can quickly grasp developments in modern AI systems and infer how to maximally leverage them to understand the world. You can start with an open problem – determining how likely it is that a given type of information produced by an LLM is true, for example – research an innovative solution, and build great evals to convince us that it’s working. You’re excited that your research plays a key role in producing maximally trustworthy and accurate information on questions of the utmost importance.
Benefits
We are fully remote. SF and London are our primary in-person locations, and we travel to meet quarterly. We have flexible hours.
Our salary range is $100,000 to $185,000, varying by country and seniority. Along with benefits and equity, you should expect a total compensation package that is near the top for any seed-stage startup, with large room to grow.
We also offer higher-than-average equity.
We offer full benefits, varying based on your country, including: healthcare, paternity/maternity leave, equipment, and travel expenses.
Similar Jobs
What We Do
At FutureSearch, we build AI to understand AI, in a virtuous loop. When you’re making million-dollar decisions, no LLM system today, from grok-3 to DeepSeek to OpenAI Deep research, is anywhere close to reliable enough to use. At FutureSearch, our research services are powered by our elite human forecasters using our sophisticated in-house AIsystems. Our ability to predict the future of AI better than the best humans — comes from being the only team in the world that deeply understands both modern AI, and how the best human forecasters think.
