Build where infrastructure meets intelligence

We use machine learning and AI to solve hard problems in high-stakes, adversarial environments. Competing here means pairing smart models with fast systems and thinking that can keep up with a market that never stops moving.

This is a place for people who want their work to matter. New ideas meet real infrastructure, clean data, and fast feedback. We focus on what works, why it works, and how to make it better. We build for extreme complexity, scale, and longevity. We’re hiring people who want to do the same.

Explore AI/ML Roles

“Traders figure out how to profit tomorrow, but research is what lays the foundation to win two years from now. We strive to understand where Jump should be before the market does.”

Lucas Baker

Head of LLM R&D

What we work on

Petabyte-scale data. Nanosecond decisions. Creativity supported by rigor.

Machine learning is deeply integrated across trading, research, and core infrastructure. Our techniques range from classical models, executed with extreme scale and precision, to frontier-level work in deep learning, RL, LLM, and generative modeling.

Our core focus is finding alpha in high-volume, low-latency environments with low signal-to-noise ratio. This demands an adaptable modeling process, carefully refined data, rigorous signal validation, and seamless execution.

10,000+

Compute Nodes

5M+

Simulations Run Daily

Trading Across 100s
of Global Exchanges

A few examples:

Time-series forecasting in non-stationary, adversarial environments
Real-time inference on petabytes of market and alternative data
Massive HPC grid with thousands of GPUs and blisteringly fast I/O
LLM agents and assistants, served through API and HPC
NLP analysis and signal generation from unstructured data
AI-enabled software and trading tools to accelerate speed of research

How we work

Ambitious projects. Tight loops.

Researchers, engineers, and traders work side by side to take ideas from concept to production, combining a scientific mindset with fast, focused iteration. Some of our ML and AI experts sit directly with trading teams to specialize in specific markets. Others work centrally to build scalable systems, high-performance infrastructure, custom foundation models, and LLM agents integrated across our tools and data.

< 24h

Model Deployment to Trader Feedback

50+

AI Tools
Used Daily

75%+

Firm-wide LLM
Weekly Usage

The nature of our problems demands:

Evolving alpha: Any single source of edge erodes over time as the market grows more efficient. Adapt or lose.
Adverse selection: Every time you trade, it’s because someone thought you were wrong. “Good on average” doesn’t cut it.
Extremely low signal-to-noise: It’s incredibly easy to fool yourself. Data not pristine? You’ll fit the noise, but not the signal.
Hard performance constraints: It’s not just latency. Can your model process every symbol under tight hardware constraints? Let’s find out.
Massive, unstructured datasets: It’s easy to demo your idea in a notebook. What’s valuable is making it 99.9% accurate on real data, parsed on the fly.

Everything we build gets tested, deployed, and improved—fast.

Who Thrives Here

Problem solving > pedigree

We hire PhDs, engineers, researchers, and scientists from all backgrounds. If you have experience in finance or trading, great. If not, we’re more than happy to teach you. While we prioritize candidates with strong quantitative intuition, engineering skills, and research experience in machine learning or AI, our only requirements are intellectual curiosity, rigor, and a drive to build the future of quantitative trading.