AI engineer / researcher (LLM agent focus)

About Mantic Mantic is an AI startup based in London. Mantic's mission: solve judgmental forecasting to radically improve decision-making. Judgmental forecasting is about making predictions in the absence of good statistical models, often on the topics of geopolitics, business & economics, policy, and technology. Human superforecasters use reasoning and research, and we're building an AI system that does the same. We have an unusually strong technical team, coming from previous work at Google DeepMind, Citadel, Foresight Data Machines, the Bank of England, Goldman Sachs, Palantir, the VGG group at the University of Oxford, and the University of Cambridge. According to Polymarket , Mantic is "by far the best performing AI forecasting system that we are aware of". Mantic placed 4th / 500+ people in the most recent Metaculus Cup, a leading forecasting competition. ( The Atlantic , The Guardian .) The role We're looking for a mission-driven candidate to help us push the frontier of forecasting accuracy. The role will primarily involve working with LLM agents. You'll find improvements to scaffolding and run experiments, evaluating on internal benchmarks. The typical project length will be short (e.g. 1 week). Example activities: Making new benchmarks Experimenting with new tools for letting LLMs express probability distributions Building out deep research agents Analyzing why we made bad predictions on certain questions Your title will be Member of Technical Staff. Essential: Strong work ethic. Strong coding ability. Experience working on research projects. Comfortable with statistics. Desired: Experience with LLM agent frameworks. Experience working in quant finance, an AI lab, an AI startup, or academic research on LLMs. Strong academic results. Evidence of exceptional ability or achievement (from any domain). Location: In-person, 5 days per week. The Mantic offices are in Camden, on the King's Cross side. We will consider visa sponsorship for exceptional candidates. Mantic has a skilled worker sponsorship licence. Benefits: Generous equity. Competitive salary. Lunch. Join a world class team, working together on one of the world's most exciting problems.