Machine Learning Researcher

About the Role

Machine learning is not a feature at Uncharted Network — it is the core of our trading operation. As a Machine Learning Researcher, you will design and train the deep learning models that power our next-generation strategy layer, working on problems that sit at the intersection of large-scale model training and financial market prediction. Our environment is unusual: you will have a tight feedback loop between model iteration and live trading performance, so you will know quickly whether your ideas work. We draw from the full ML landscape — from sequence models and attention-based architectures to reinforcement learning and classical methods — and we are not ideological about tools.

Responsibilities

• Train and evaluate deep learning models on financial time-series and alternative datasets for live strategy deployment
• Explore novel modelling approaches including sequence models, reinforcement learning, and ensemble methods
• Collaborate with ML Engineers to bring research-ready models into production training and inference pipelines
• Design rigorous experiments, track results, and produce clear documentation of model behaviour under regime changes
• Hire, mentor, and conduct technical reviews with fellow researchers and engineers
• Engage with academic literature and conferences to bring state-of-the-art techniques into the firm

Requirements

• Demonstrated research experience in empirical machine learning, with publications or equivalent project depth
• Strong mathematical foundations: linear algebra, optimisation theory, probability theory, and statistics
• Fluency in Python and at least one major ML framework (PyTorch, JAX, or TensorFlow)
• Experience designing and running large-scale experiments with reproducible, well-documented results
• Ability to apply logical and mathematical thinking across financial and non-financial problem domains
• Excellent written communication; able to produce research memos for the investment committee

Nice to Have

• Experience with non-stationary datasets, distributional shift, or multi-agent environment modelling
• Prior work on financial time series, tick data, or order book prediction
• Familiarity with distributed training across multi-GPU or multi-node GPU clusters