AI Lab - Research
OptiverAbout the Role
The AI Lab is a research-focused trading team exploring how advances in machine learning can be applied to complex problems in quantitative research and trading. The Lab develops and runs its own ML-driven trading strategies, owning the full lifecycle from research and experimentation through production and revenue generation. It sits within the Quantitative Strategy Group (QSG) and brings together researchers/traders, machine learning engineers, and software engineers.
As a Researcher in the AI Lab, you will:
- Design, develop, and train novel ML models and methods — including LLMs and other foundation models — drawing on advances in deep learning and sequence modeling, for deployment in production trading systems
- Research and develop new ML approaches for complex quantitative and sequential modeling problems
- Formulate hypotheses and design rigorous experiments and evaluation frameworks, using appropriate baselines, ablations, and out-of-sample testing to identify promising research directions and understand why approaches succeed or fail
- Translate ideas from research papers and theoretical work into working implementations, adapting and extending them to new problem domains
- Apply ML techniques to large-scale financial datasets, including time-series and unstructured data, to identify and evaluate potential predictive signals
- Train and evaluate models at scale, leveraging GPU and distributed computing environments as needed
- Build and leverage research tooling, including agentic and automated research workflows, to accelerate experimentation
Responsibilities
- Designing, developing, and training novel ML models and methods — including LLMs and other foundation models — drawing on advances in deep learning and sequence modeling, for deployment in production trading systems
- Researching and developing new ML approaches for complex quantitative and sequential modeling problems
- Formulating hypotheses and designing rigorous experiments and evaluation frameworks, using appropriate baselines, ablations, and out-of-sample testing to identify promising research directions and understand why approaches succeed or fail
- Translating ideas from research papers and theoretical work into working implementations, adapting and extending them to new problem domains
- Applying ML techniques to large-scale financial datasets, including time-series and unstructured data, to identify and evaluate potential predictive signals
- Training and evaluating models at scale, leveraging GPU and distributed computing environments as needed
- Building and leveraging research tooling, including agentic and automated research workflows, to accelerate experimentation
Requirements
- Strong foundations in machine learning, statistics, optimization, and experimental design
- Demonstrated ability to conduct independent ML research, from developing hypotheses through implementation, experimentation, and evaluation
- Deep understanding of modern deep learning architectures, particularly transformers, foundation models, sequence models, and/or state-space models
- Experience developing and training models rather than primarily applying or integrating existing models
- Strong empirical judgment and the ability to understand why an approach is or is not working and determine the next research direction
- Strong programming skills, particularly Python, with experience in frameworks such as PyTorch or JAX
- Experience training and evaluating models in GPU-based computing environments
- Experience in one or more of the following would be especially valuable: developing and training large-scale foundation models or LLMs
- Experience in one or more of the following would be especially valuable: large-scale sequence or time-series modeling
- Experience in one or more of the following would be especially valuable: transformers, efficient attention, SSMs, Mamba, RWKV, or related architectures
- Experience in one or more of the following would be especially valuable: reinforcement learning or sequential decision-making
- Experience in one or more of the following would be especially valuable: self-supervised, representation, or generative learning
- Experience in one or more of the following would be especially valuable: pre-training, post-training, or novel model architecture research
- Experience in one or more of the following would be especially valuable: distributed training and large-scale GPU workloads
- Experience in one or more of the following would be especially valuable: quantitative research, financial markets, or high-frequency data
- Experience in one or more of the following would be especially valuable: CUDA, Triton, custom kernels, or ML performance optimization
- Experience in one or more of the following would be especially valuable: agent-assisted or automated research methodologies
What You'll Get
- $200k base salary
- The opportunity to work alongside best-in-class professionals from over 40 different countries
- 401(k) match up to 50%
- Comprehensive health, mental, dental, vision, disability, and life coverage
- 25 paid vacation days alongside market holidays
- Extensive office perks, including breakfast, lunch and snacks, regular social events, clubs, sports leagues and more