AI Scientist

Role overview

Build machine learning models and AI agents that generate investigative intelligence from blockchain and related data. The work spans research, evaluation, and production systems, with a focus on discovering patterns that help investigate illicit activity and threats. You’ll collaborate with specialists who provide domain expertise and high-confidence labels, and move promising ideas quickly toward customer use.

Responsibilities

- Train and evaluate machine learning models, including foundation models, using blockchain data. - Develop AI agents and, where useful, adapt language models to produce investigative intelligence. - Explore signals across network, on-chain, and protocol data to uncover new connections and patterns. - Deliver production-ready models, agents, and data pipelines with scalable designs and rigorous evaluation. - Take research ideas from early prototypes to customer-facing systems, incorporating feedback and measuring impact. - Coordinate across teams and priorities while maintaining quality and momentum.

Requirements

- At least five years of hands-on machine learning experience, including applied modeling, research, experimentation, or evaluation. - Strong Python and SQL skills, plus experience delivering production machine learning systems. - Experience with distributed data tools such as Spark, BigQuery, or Airflow. - A quantitative background in computer science, statistics, engineering, mathematics, physics, or equivalent experience; advanced study or research experience is a plus. - Ability to investigate ambiguous problems through clear hypotheses, iterative experiments, and independent ownership. - Strong written and spoken technical communication in English, including async collaboration in a remote setting.

Nice to have

- Experience with language models, AI agents, or fine-tuning. - Interest or experience in blockchain, cryptocurrency, fraud, money laundering, or financial crime. - Familiarity with AI-assisted coding tools.

Benefits and work setup

- Fully remote, globally distributed team; the role requires at least four hours of overlap with Pacific business hours, with much team collaboration typically scheduled in the morning Pacific time. - Work includes short-cycle prototypes and rapid iteration based on customer feedback.