Deep Learning Quantitative Researcher

About the role

Millennium is seeking a Deep Learning Quantitative Researcher to design and build core deep‑learning pipelines that drive applied quantitative alpha research. The role involves end‑to‑end model development, from data preparation through production deployment, within a fast‑moving, low‑signal environment.

Key responsibilities

• Design and implement deep‑learning pipelines for quantitative research, covering data preparation, distributed training, evaluation and production deployment.
• Lead the research agenda using applied deep‑learning techniques, owning problem formulation, model design, training, validation and performance attribution.
• Maintain rigorous research discipline with strict out‑of‑sample hygiene, leakage prevention and honest benchmarking against simpler baselines.
• Serve as the central deep‑learning expert, advising on architecture selection, training diagnostics, reviewing model designs and setting evaluation standards.
• Standardize training and inference interfaces to enable seamless model fitting across teams and systems.

Required profile

• Top‑tier academic background (top‑20 global university) with PhD‑level training in Computer Science, Engineering, Physics, Mathematics or Statistics.
• 3–5 years of professional experience applying deep learning to large‑scale problems, preferably in quantitative finance or a leading AI/technology company.
• Proven end‑to‑end ownership of a deep‑learning model lifecycle on a production system or published research line.
• Strong foundations in statistics, optimization and machine‑learning theory.

Required skills

• Python programming.
• Modern deep‑learning frameworks (e.g., PyTorch, TensorFlow).
• Distributed/multi‑GPU training, mixed‑precision and throughput optimization.
• Experience with large‑scale datasets, columnar formats and streaming data loaders.
• Experiment‑management tooling, hyper‑parameter optimization and reproducible research environments.
• Working knowledge of C++ or CUDA‑level optimization (optional).
• Familiarity with LLM tooling as a research accelerant (optional).