ML Research Scientist -Deep Learning & Transformer Architectures
Core
Design, implement, and train a custom decoder-only Transformer model from scratch for next-token prediction of intraday price movements using tokenized market data.
Role type
Senior IC ML Research Scientist (Deep Learning & Transformer Architectures)
Builds
Proprietary Transformer-based models for systematic equities trading
Domain
Financial technology / Quantitative trading / Deep Learning
Deliverable
production ML models
Required skills
Transformer architecture design, PyTorch (custom modules, mixed-precision, multi-GPU), tokenization strategies, attention mechanisms, Python, C++, mathematical foundations (linear algebra, probability, optimization), large-scale model training (100M+ parameters)
Preferred skills
Financial time-series forecasting, tokenization for non-text data, published research in top ML venues, market microstructure knowledge, model compression and inference optimization
Technologies
PyTorch, Python, C++