Senior Data Scientist: Info-Driven Trading & Validation

TWG AI; Santa Monica (CA); On-site; USD 190,000 - 290,000; Full time

Benefits offered by this job

Medical benefits

Bonus

Job summary

TWG AI is seeking a Senior Data Scientist to own the detection and validation components of a regulated surveillance system. You will separate public from non-public information timing, and ensure detectors produce real signal with on-chain and ground-truth resolutions.

This role blends research depth with practical, audit-ready implementation. You will mentor junior colleagues while expanding validation across markets, building trust in model outputs and improving labeling processes in a

Qualifications

• Strong empirical and statistical background with hypothesis testing and backtesting.
• Experience with weak/noisy labels and human-in-the-loop labeling systems.
• Quantitative finance or empirical research background is a strong plus.
• Experience with sequential models (transformers/RNNs) for behavior/transaction sequences.
• Production ML in regulated, audit-facing work.

Responsibilities

• Build detection logic for insider and information-driven trading across venues.
• Design and run rigorous validation and backtesting against ground truth.
• Own labeling quality and continually improve the DeFi label corpus.
• Grow the validation work into an audit-ready assurance capability.
• Mentor junior team members contributing to validation and analysis.

Skills

Empirical statistics

Labeling systems

Quantitative finance

Sequential modeling

Production ML in regulated env

Education

PhD or equivalent industry experience

Job description

TWG AI is seeking a Senior Data Scientist to own the detection and validation components of a regulated surveillance system. You will separate public from non-public information timing, and ensure detectors produce real signal with on-chain and ground-truth resolutions.

This role blends research depth with practical, audit-ready implementation. You will mentor junior colleagues while expanding validation across markets, building trust in model outputs and improving labeling processes in a