Senior Machine Learning Engineering Manager

Signifyd

2d•$200,000 - $245,000•Remote

About The Position

At Signifyd, we help merchants confidently grow their businesses by building trusted relationships with their customers. Our advanced technology, combined with a team genuinely invested in our clients’ success, creates frictionless shopping experiences, approving more good orders, protecting revenue, and keeping customers happy. Trusted by thousands of leading merchants across more than 100 countries, we securely process billions of transactions each year. Our people are the heart of everything we do, driving our mission forward with commitment, empathy, and creativity. Join us on our mission to empower confident, fraud-free commerce by helping online retailers provide superior customer experiences and eliminate fraud. The Signifyd AI Lab (SAIL) builds the ML products behind Signifyd's fraud and risk decisions. We improve the predictive performance of the models that decide e-commerce transactions at scale, we scale the ML capabilities of our Risk organization, and we push into new markets and problem spaces that expand the market Signifyd can sell to. Every space in this department is a mix of experimentation, code, and statistics. We don't create walls between the people who have the ideas and the people who build them. The team splits its time between near-term continuous model improvements and longer-horizon innovation bets to improve the company’s capabilities in 2027 and beyond. These bets surface from the ground up in an environment where we believe those closest to the problems are best placed to understand how to solve them. We’re hiring a manager to lead one of the teams in this department.

Requirements

Roughly 5+ years in machine learning, data science, or ML-adjacent software engineering, including at least 3 years of people management — guiding career development, addressing conflicts, and building a healthy, high-performing team.
Genuine depth in at least one of engineering and applied statistics, and real working competence in the other. We are not hiring a manager of analysts, and we are not hiring a manager of a pure software team. Our engineers train production models that decide serious traffic, and we expect their manager to be able to engage with that work at a technical level.
Demonstrated ability to lead work under real uncertainty: setting a direction when the answer isn't known yet, changing course when evidence says to, and communicating both without eroding your team's confidence.
Excellent written and verbal communication. Much of our decision-making happens in documents, and we expect managers to write well.
Autonomy in recognizing priorities and evaluating the impact of outcomes, and comfort working without close supervision in a fast-moving environment.
Commitment to quality. You take pride in work that excels in correctness, reproducibility, and reliability, and you set that standard for your team.

Nice To Haves

You are a hands-on Player-Coach who thrives in ambiguity—where the roadmap is a set of hypotheses, and the answer to "will this work?" is "we'll know in three weeks."
Technical Credibility (The "Player"): You stay close enough to the work to have a grounded opinion. You read the code, inspect evaluation pipelines, and can immediately tell the difference between a statistical result that will hold up in production and one that just happened to look good on a single test window.
Leadership & Rigor (The "Coach"): You hold a high bar for evidence without becoming a bottleneck to experimentation. You mentor engineers to own their code quality, and you translate complex ML performance metrics into clear business outcomes for Risk leadership.
Executive Judgment: You know how to balance research bets against quarterly delivery, disagree and commit when decisions are made, and build an environment where well-documented negative experimental results are celebrated as real progress.

Responsibilities

Lead and grow the team
Guide career development, manage conflicts, and nurture a positive work environment.
Develop career plans with team members, provide guidance on skill development, and follow up on their evolution.
Engage in regular 1:1s, give constant feedback, and create a safe environment for open discussion — including the discussions that follow an experiment that didn't work.
Set clear goals, mentor the team, and foster a collaborative environment across a geographically distributed organization.
Encourage a culture of learning and improvement, provide technical guidance, and support team members in both technical and soft skills.
Conduct technical and hiring-manager interviews, train the team on interviewing techniques, and help us keep raising the bar as we grow.
Identify and address gaps in team capabilities and processes to enhance team efficiency and success.
Partner with your tech leads, who own and drive the technical roadmap for their areas. Your job is not to be the sole source of ideas — it is to pressure-test them, sharpen them, make sure the strongest ones get resourced, and make sure the people generating them have the room and the support to do it. When you do bring an idea, you bring it as a peer in the technical conversation.
Make the calls the roadmap can't make for you: which hypotheses get compute and headcount, which get another iteration, and which get a clear, documented "no." A well-run negative result is a real outcome, and we treat it as one — but only if it's declared, written down, and learned from.
Manage the trade-off between a committed improvement target you must hit this year and research bets that may not pay off for several quarters. You will re-cut that budget as evidence arrives, and you'll be able to explain the reasoning to both your team and your stakeholders.
Bring rigor to how the team decides something worked. Offline results have to predict online behavior; a strong point estimate on a single evaluation window is a starting point, not a conclusion. You will be the person asking whether the improvement survives a rolling evaluation, whether it's already captured by a change we shipped last month, and what would have to be true for it to be wrong.
Own delivery on a cadence. Independent experimental workstreams have to converge into a release candidate, get evaluated end to end, and ship — including the hard call to leave a workstream out of a release when it isn't carrying its weight.
Work directly with our Risk partners as your primary stakeholders. Our commitments to them are explicit, measured, and written down; we deliver model performance, and they own thresholds, rules, and how decisions are applied to merchants.
Operate with a high degree of autonomy. Our direction comes from measured performance against those commitments and from what our own experiments tell us, not from a product backlog handed to the team. You are expected to know what your team should be working on and to defend it, rather than wait to be told.
Partner with our platform and infrastructure engineering teams on the feature systems, training pipelines, and experimentation tooling your team depends on — and be clear about where the boundary sits between what SAIL should own and what belongs to Engineering.
Represent your team's results to a broad audience: engineering leadership, Risk leaders, and the wider company.