Deep Learning Quantitative Researcher

Position Overview

Millennium is seeking a Deep Learning Quantitative Researcher to join their team in Singapore. The ideal candidate will have exceptional academic qualifications, advanced training, and robust experience in applied deep learning within quantitative finance or related AI/technology fields.

Key Responsibilities

• Develop and maintain the core deep learning infrastructure supporting quantitative alpha research, covering the entire pipeline from data preparation and distributed training to evaluation and deployment.
• Lead significant research initiatives focused on applied deep learning methods, managing the entire empirical cycle including problem statement, model design, training, validation, and performance attribution.
• Maintain stringent research standards in a challenging low signal-to-noise environment by ensuring out-of-sample integrity, preventing data leakage, and conducting honest benchmarking against simpler models.
• Serve as the primary deep learning expert within the firm by providing guidance on architecture choices, training diagnostics, model reviews, and setting evaluation and promotion standards.
• Enable smooth integration of model fitting and computation across teams and systems by standardizing training, inference interfaces, and reusable components.

Qualifications and Experience

• 3 to 5 years of professional experience applying deep learning to extensive problems, preferably in quantitative finance. Exceptional PhD-level research and practical experience at a leading AI or technology firm are acceptable substitutes.
• Demonstrated full ownership of deep learning model lifecycles, with at least one major production system or published research project.
• Deep proficiency in Python and modern deep learning frameworks.
• Experienced in large-scale model training including distributed/multi-GPU setups, mixed precision computing, and performance optimization.
• Strong foundational knowledge in statistics, optimization, and machine learning theory.

Technical Competencies

• Expertise with modern deep learning architectures and the discernment to prefer simpler models when appropriate.
• Mastery of techniques for learning in low signal-to-noise scenarios, including regularization, ensembling, and robust validation ensuring out-of-sample reliability.
• Experience handling large datasets efficiently, including the use of columnar data formats, streaming loaders, and point-in-time correct dataset construction.
• Familiarity with experiment management tools such as tracking systems, hyperparameter tuning, and reproducibility frameworks.
• Additional skills such as knowledge of C++ or CUDA optimization and experience with large language model tooling are advantageous.

Desired Personal Attributes

• High research standards with a disciplined approach to experimentation and decisive elimination of unsupported ideas.
• Strong collaborative spirit to build effective partnerships between research and engineering teams.
• Unwavering ethical considerations when managing sensitive data and models.
• A commitment to continuous learning and adaptation to evolving technologies and methodologies.
• Excellent communication skills to articulate model functions and uncertainties to both technical and non-technical stakeholders.

Application Instructions

Interested candidates should submit their resumes referencing REQ-30088 to the provided recruitment contact email.