ML Engineer

As a MLE you'll join our Product Innovations team and work across the full applied ML stack - deploying models, building the evaluation systems that tell us whether they actually work, and making the data and infrastructure decisions that turn experimental data science into cost-efficient products. You'll partner closely with our Data Science and Engineering teams on our vector embeddings ecosystem, ground truth pipelines, model evaluation, and the pre/post-processing decisions that determine product quality. This is a production focused role, with some research opportunities. You'll be the engineer who makes sure our ML systems - both traditional NLP and embedding models and our LLM-powered features - work reliably at scale (millions of records per day), are continuously evaluated against ground truth, and improve over time.

Responsibilities

Deploy and monitor ML systems in production, from classical NLP and embedding models to LLM-powered features - where "production" means millions of records per day; Own the evaluation stack - golden datasets, "model-as-a-judge" frameworks, inter-annotator agreement, and regression tests that gate releases; Build and maintain our vector embeddings ecosystem and the retrieval, classification, and similarity patterns that sit on top of it; Partner with Data Science on annotation workflows, PII scrubbing, and ground-truth pipelines; Improve our MLOps foundations - versioning, observability, drift detection - so the rest of the team can ship faster; Translate fuzzy product problems into measurable AI features with clear success criteria

Requirements

4–7 years of professional software or ML engineering experience, including 2+ years shipping ML systems to production; Strong Python; comfort with the modern data/ML stack; Hands-on experience deploying and monitoring models in at least one major cloud (AWS or GCP); willingness to learn the other; Production experience with NLP or ML systems - classification, NER, embeddings, ranking, similarity, or LLM-powered features (most candidates have done some mix of traditional ML and LLM work; we care that you've shipped, not which camp you came up in); Practical experience with evaluation for ML or LLM systems - golden datasets, model-as-a-judge, IAA, precision/recall, or equivalent. You don't need to have built one from scratch, but you should know why they matter and how to improve them; Collaborative communicator - you work well alongside data scientists and engineers, and can clearly explain ideas, requirements, and tradeoffs to non-technical stakeholders

Nice to have

Experience with vector databases or retrieval systems at scale; Experience with managed ML services on AWS (SageMaker) and/or GCP (Vertex AI); Annotation workflow experience (Label Studio, Scale AI, or similar) and a point of view on inter-annotator agreement; Familiarity with PII scrubbing patterns and privacy-by-design data handling; Open-source contributions, blog posts, or talks on LLM/embedding production work

Benefits

People: work with talented, collaborative, and friendly people who love what they do; Guidance: utilize our learning platform to fully get the training and tools you’ll need to become successful here from your first day with us; Surprise meal stipends: work from home can’t stop the enjoyment of someone else making a meal for you; Work/life harmony: vacation, floating and set holidays, wellness allowance, and paid parental leave; Whole Health Package: medical, dental, vision, life, disability insurance, and more; Work from home stipend: to assist you in setting up a home office that works for you (or buy a new dog leash - your choice!)