About the job
About Middesk
At Middesk, we simplify the way businesses connect and collaborate. Since our inception in 2018, we have revolutionized business identity verification by replacing tedious, manual processes with instant access to comprehensive, up-to-date information. Our innovative platform empowers companies across various sectors to confidently verify business identities, expedite customer onboarding, and mitigate risks throughout the customer lifecycle.
Originating from Y Combinator and backed by notable investors such as Sequoia Capital and Accel Partners, Middesk has recently earned a spot on the Forbes Fintech 50 List and has been recognized as a leader in business verification by digital identity strategy firm Liminal.
The Role
We are on a mission to develop AI-driven applications that enhance customer workflows, particularly in the realm of business onboarding. Leveraging our proprietary identity data and extensive domain expertise, we are uniquely positioned to broaden our suite of AI-powered solutions that fuel sustainable growth.
We are seeking a hands-on applied Machine Learning expert to establish the technical foundation for these initiatives. The ideal candidate will have experience deploying external-facing models in the risk and fraud sectors and will be familiar with the complexities of imbalanced data, limited labels, and evolving behaviors. This is a highly technical and influential role that will shape our ML design, development, and scaling efforts at Middesk.
What You'll Do:
Develop risk and fraud ML applications: Deliver production ML models in fraud detection, trust and safety, Know Your Business (KYB), and compliance, with a measurable influence on customer workflows.
Address challenging data issues: Work on classification tasks characterized by extreme class imbalance, sparse signals, and “cold start” label challenges.
Innovate in feature engineering and labeling: Implement graph-based methodologies, weak supervision, LLMs, and AI agents to enhance signal extraction and automate the labeling process.
Establish foundational ML infrastructure: Collaborate with the platform engineering team to design feature services, model training pipelines, model serving standards, and orchestration to scale multiple ML applications.
What We’re Looking For:
A minimum of 7 years of applied ML experience, with demonstrable impact in risk, fraud, trust and safety, compliance, or related high-stakes domains.
A proven record of deploying ML models from research to production in client-facing products.
Expertise in classification challenges, such as imbalanced labels, sparse signals, and cold start issues.