SlipstreamJobs tracks this role from the company's public career site. Apply directly on the employer's site.
Salary: CAD 299,300 - 372,600 / annual
Babylist is seeking a Staff Machine Learning Engineer to own personalization across the platform. As a Staff MLE, you will set direction for personalization as a domain, owning the models behind the homepage feed, add-next recommendations, search personalization, and foundational representations that multiple teams build upon. You will sequence technical bets over the next 1-2 years and make critical modeling and architecture decisions that span teams.
Key responsibilities include:
- Taking fuzzy business problems from initial concept through production, staying accountable for customer impact
- Building custom embeddings from raw data—domain-specific representations beyond off-the-shelf models—and owning them as multiple surfaces adopt them
- Making modeling and architecture calls that span teams and are expensive to reverse
- Owning the full ML lifecycle: orchestration, deployment, monitoring, and retraining loops
- Setting standards for how personalization integrates with AI; building evals to catch confident failures before shipping
- Partnering with product, design, and data as a peer from problem definition onward
- Coaching Senior engineers through hard and ambiguous technical decisions
Current focus areas include building foundational embeddings shared across feed, recommendations, and search; resolving customer identity across registry, shop, and health; and determining registry recommendations with live experiment validation.
Babylists serves 10 million gift-givers annually, generated $750M+ in revenue in 2025 (45% YoY growth), and has been profitable for 8 years while remaining independent. The engineering team is ~65 people, so your work stays visible and your scope remains wide. The company is remote-first across US and Canada, with in-person team gatherings twice yearly. Teams are small pods of 3-5 engineers. You will work closely with product, design, data, and business partners, and regularly engage with customers through user interviews, session recordings, and support interactions.
The tech stack includes Rails, React, TypeScript, MySQL, Snowflake, dbt, Airflow, Weaviate, AWS SageMaker, MLflow, and AI tooling like Claude Code and Devin. The architecture is intentionally simple—one Rails monolith with few moving parts—to enable speed and allow AI to reason about the whole system.
REQUIREMENTS:
- Multiple years of shipped production ML with strong, loosely-held opinions
- Demonstrated ability to pick up ambiguous problems and move before the full picture is clear
- Proven track record changing how a team builds with AI, with the new approach sticking
- Built recommender systems or personalization reaching real users at scale; can point to measurable impact
- Deep expertise in Python ML ecosystem (pandas, scikit-learn, XGBoost, PyTorch)
- Fluency across the full ML lifecycle: orchestration, training, monitoring—not just model 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 that multiple surfaces depend on
- Zero-to-one capability: define problem space, architect from scratch, own end-to-end
- Curiosity: spot problems before they're filed, push your own ideas until they ship
