About this role
Summary: Own end-to-end ML systems for fraud and abuse detection across Stripe's global network. Design, build, and deploy production models using modern approaches like transformers and graph intelligence to protect businesses while ensuring low latency and high reliability.
Who we are
### About Stripe
Stripe is a financial infrastructure platform for businesses. Millions of companies—from the world’s largest enterprises to the most ambitious startups—use Stripe to accept payments, grow their revenue, and accelerate new business opportunities. Our mission is to increase the GDP of the internet, and we have a staggering amount of work ahead. That means you have an unprecedented opportunity to put the global economy within everyone’s reach while doing the most important work of your career.
### About the team
The Transaction Risk team builds the machine learning systems behind Stripe Radar, helping businesses prevent fraud and abuse while letting legitimate customers through. We combine intelligence from across Stripe’s global network with real-time models and AI agents to protect businesses across the customer lifecycle, from signup and free trials to payments and ongoing usage, with a growing focus on the emerging threats facing AI companies.
The team is growing, and expanding into multiple new areas. We are building abuse protections across the full customer life-cycle and solving emerging token fraud problems to make Stripe the default fraud and abuse solution for AI.
What you’ll do
You will own major ML systems and initiatives from problem definition to production. You'll find where ML can make the biggest difference, design the approach, build and ship models and the systems around them, and keep improving them as data, users, and attackers change.
Your work may include protecting businesses from abuse across the customer lifecycle, building adaptive and customizable fraud systems for Stripe's largest users, or extending risk intelligence to 50+ global payment methods. Across all of these, you'll use modern approaches, such as transformer models over payment data, graph intelligence across consumers, and LLMs that detect emerging attacks.
You will have opportunities to ship 0→1 products, and to improve mission-critical ML models operating at Stripe scale, often driving several efforts in parallel and setting the technical bar for the team.
### Responsibilities
- Lead the design and delivery of ML projects end to end, from framing the problem and sizing the opportunity through deployment, monitoring, and iteration.
- Build, train, evaluate, and deploy models that detect fraud and abuse across Stripe’s global network, and own the production systems they run on.
- Investigate emerging fraud and abuse patterns, decide where ML can have the most impact, and quickly turn those insights into shipped protections.
- Design features, datasets, and evaluation strategies for noisy, delayed, and adversarial environments, and set the bar for experimental rigor.
- Make sound architecture and modeling tradeoffs between model quality, legitimate-user impact, latency, reliability, and operational cost.
- Be the go-to expert for the systems you own: lead incident response for model regressions, reduce operational toil, and make your systems easy for others to build on.
- Partner with engineering, product, data science, risk experts, and leading Stripe users to turn modeling improvements into differentiated Radar products.
- Turn recurring problems into reusable components, tools, and AI-assisted workflows that speed up the whole team.
- Mentor engineers, give thoughtful design and code reviews, and help new teammates ramp up quickly.
Who you are
You are an experienced ML engineer who combines deep modeling expertise with strong engineering judgment. You've owned ML systems in production, shipped multiple meaningful improvements, and made the people around you more effective. You move fast without breaking things, you're comfortable leading through ambiguity, and you're excited to work at the frontier of payments and AI, turning advances in machine learning into production systems that address new forms of fraud and abuse.
We’re looking for someone who meets the minimum requirements to be considered for the role. If you meet these requirements, you are encouraged to apply. The preferred qualifications are a bonus, not a requirement.
### Minimum requirements
- 5+ years of experience developing, evaluating, and deploying machine learning models in a production environment.
- A track record of leading ML projects from problem definition to production, with measurable impact.
- Deep foundations in machine learning, statistics, model evaluation, and experiment design.
- Strong programming and software-engineering skills, with experience writing tested, maintainable code for production ML applications.
- Experience designing and operating production ML systems, including several of: large-scale data and feature pipelines, distributed model training, low-latency inference, model deployment, or production monitoring.
- Experience mentoring engineers or technically leading projects with multiple contributors.
- Ability to investigate ambiguous problems, reason from data, and choose pragmatic technical approaches.
- Ability to communicate technical ideas and results clearly and collaborate effectively across disciplines.
### Preferred qualifications
- Experience applying ML to fraud detection, integrity, trust and safety, identity, cybersecurity, or another adversarial domain.
- Experience developing or productionizing modern deep learning, graph-learning, or LLM-based approaches.
- Experience building and optimizing real-time, low-latency ML systems at scale.
- Experience owning the operational health of critical ML systems, including on-call, incident response, and migrations.
- Experience rapidly building 0→1 products or working directly with customers.
- Experience building AI-assisted tooling or agents that measurably increased a team's velocity.