About the team
The Predict team builds Alloy’s real-time machine learning systems at scale. Our immediate focus is on fraud detection, where we believe machine learning can simplify and accelerate decision-making in ways traditional rule-based systems can’t. Managing rules and policies to detect fraud is complex and constantly evolving; we use ML to make it smarter, faster, and more adaptive. Our approach is identity-centric, combining signals from a wide range of data sources to build a comprehensive understanding of risk.
You will work on advancing our core models while also partnering directly with customers to drive strong outcomes from fraud studies.
Alloy operates in a hybrid work environment. We look to foster collaboration and community by having our local employees onsite three days a week.
What you'll be doing
- Contribute to the design, training, and evaluation of machine learning models that power Alloy’s fraud detection capabilities.
- Develop testing plans, metrics, performance reports, and translate findings into actionable recommendations.
- Support production ML workflows, including feature generation, model training, and monitoring, to ensure models remain accurate and reliable at scale.
- Document findings and communicate insights to internal teams, contributing to shared learning and continuous improvement.
- Maintain up-to-date model documentation and support Alloy’s model governance processes to ensure transparency and compliance.
- Stay current with industry trends in applied ML and fraud detection, and contribute to Alloy’s mission of making financial services safer and more accessible.
Who we’re looking for
- Always building with end-solution in mind.
- Able to communicate complicated concepts to a non-technical audience without diluting the complexity of the work.
- Able to build strong cross-functional relationships within Alloy.
- Naturally curious with a knack for asking tough questions.
- Solid understanding of core ML concepts such as supervised learning, feature engineering, and model evaluation.
- A team player. You believe that big things happen when the right people are working together.
- A fast learner
- Humble. Mistakes happen and owning them helps us learn and move on quickly
You have:
- 8+ years as an individual contributor in Applied Fraud Research, Data Science, or Machine Learning with a proven track record in a “Solutions” or client-facing capacity.
- Expertise in working with highly imbalanced datasets and building production-grade Machine Learning models with specific interest in tree-based models.
- Advanced proficiency in scripting languages like Python and querying languages like SQL
- Proven ability to wrangle and think thoughtfully about data at scale (processing billions of records).
- Experience developing metrics and dashboards.
- Able to communicate their findings effectively to technical and nontechnical members
- Someone who embodies our shared Alloy values: be bold, get scrappy, collaborate, and celebrate our differences.
- You have experience in a highly analytical role in fast-paced environments
- Must be local to Greater New York City.
Nice to Haves:
- Previous experience in financial fraud detection.
- Advanced Degree (Masters or PhD) in a quantitative field.
- Experience managing the end-to-end lifecycle of a technical pilot or Proof of Concept.
- Experience with BI tools like Looker.
- Experience with modeling on graph structures.