Staff Machine Learning Engineer

Babylist

10d•$280,200 - $348,840•Remote

About The Position

As a Staff Machine Learning Engineer at Babylist, you will own personalization and decide where it goes. Millions of families depend on what we build. You will be responsible for determining what to model, how it should work, and whether what shipped actually helped. You will still be involved in coding for genuinely hard problems like embeddings, ranking, and systems that don't exist yet, while agents handle the volume. The company is expanding beyond the registry into areas like the financial side of raising a child, maternal health, education for new parents, and community. Personalization is key across all these areas, including the homepage feed, recommendations, search, the underlying platform, and the AI already shipped to families. You will own a significant part of how personalization works, with an open roadmap and the ability to influence which bets are made and what projects you take on next.

Requirements

Shipped production ML for enough years to have developed strong opinions, held loosely.
Ability to pick up an ambiguous problem and start making progress without a complete picture.
Proven track record of changing how a team builds with AI, with the new methods sticking.
Experience building recommender systems or personalization that reached real users at scale, with demonstrable impact.
Deep knowledge of the Python ML ecosystem (pandas, scikit-learn, XGBoost, PyTorch).
Fluency across the full ML lifecycle, from orchestration to monitoring, not just training.
Ability to build custom representations from raw data, rather than relying solely on off-the-shelf embeddings.
Measure yourself by impact: customer outcomes or models depended upon by multiple surfaces.
Ability to go from zero to one: define problem spaces, architect from scratch, and own end-to-end.
Curiosity to spot problems before they are reported and drive own ideas to completion.
Experience with AI-assisted coding using tools like Claude Code, Devin, etc.
Experience with system design.
Product sense.
Culture and values alignment.

Responsibilities

Set direction for personalization as a domain, including models for the homepage feed, add-next recommendations, and search personalization, as well as foundational representations used by multiple teams.
Define the future of personalization over the next one to two years, sequence the bets to achieve it, and make technical and product decisions.
Take a fuzzy business problem from initial concept through to a production model, and remain responsible for its impact on customers.
Build custom embeddings from raw data, creating domain-specific representations that go beyond off-the-shelf models, and own them as they are adopted by various surfaces.
Make modeling and architecture decisions that span multiple teams and are costly to reverse.
Own the full lifecycle of models, including orchestration, deployment, monitoring, and the retraining loop to ensure models remain accurate in production.
Set the standard for AI-driven personalization, defining what 'good' looks like and building evaluation metrics to catch confidently incorrect models before deployment.
Partner with product, design, and data teams as a peer to shape product development from the outset.
Coach Senior engineers through complex and ambiguous technical and non-technical decisions.
Build foundational embeddings that enable all surfaces (feed, recommendations, search) to personalize from a single shared representation.
Resolve customer identity across registry, shop, and health, including their friends and family buyers, to ensure recommendations work universally.
Determine registry recommendations for each family, including ranking and the underlying model, and prove their effectiveness through live experiments.