Data Scientist
Transak is a developer integration toolkit to let users buy/sell crypto in any app, website or web plugin.
Projects
About Transak
Transak is a global web3 infrastructure services provider with registered entities in the USA (NMLS ID: 2362652), the UK, Canada, Australia, Poland, India, and Hong Kong.
Skills
About the Role
You'll take genuine end-to-end ownership of how Transak detects and prevents fraud, as part of a small, senior team. Your mandate is to stop fraud while approving as many legitimate transactions as possible — whether through deterministic heuristics, machine learning, AI agents, or whatever the problem demands. You'll work with tens of millions of transactions, ~1,000 live risk signals across 13 providers, and ~10,000 orders per hour at peak, against adversaries who adapt the moment you respond. Beyond fraud, you'll act as a data scientist for the wider business, making data accessible through the Snowflake warehouse, self-serve tooling, dashboards and internal data assistants. You'll also take on high-leverage product, growth and experimentation problems, and help advance how the company applies AI internally, from agentic coding assistants to internal copilots and novel uses of large language models. You'll have autonomy to try new models, tools and approaches, and you'll be judged on outcomes, not process or hours.
Requirements
- Strong mathematical and statistical foundations, and a genuine pull toward hard, quantitative problems
- A self-starter who identifies the important problem, scopes it, and acts without waiting to be directed
- Fluent in Python and SQL, able to take a model from idea to something that runs and ships
- Intellectually honest about uncertainty, and rigorous in evaluating your own work
- A genuine interest in crypto, payments, fraud and the data they generate
Responsibilities
- Own fraud, chargeback and transaction-risk models end to end, from framing the question to features, rules, shipping and thresholds
- Build and tune the machine-learning and signal layer that operates alongside external risk vendors
- Set the direction of where risk modelling goes next and bring the team along
- Act as a data scientist for the wider business, making data accessible through the Snowflake warehouse, self-serve tooling, dashboards and internal data assistants
- Take on high-leverage product, growth and experimentation problems as they arise
- Help advance how the company applies AI internally, from agentic coding assistants to internal copilots and novel uses of large language models