Senior Software Engineer I, Machine Learning, ML Enablement

About the role

We are looking for a Senior Software Engineer, Machine Learning to help build and operate the infrastructure that trains, hosts, and serves Etsy's machine learning models — including our internal serving platform, Barista, and a growing portfolio of predictive and generative (open source LLM) models. You will architect the systems that make model serving fast, reliable, and scalable for millions of inferences per second, enabling Applied Scientists and Engineers across Etsy to train, host, and iterate on models with confidence, from classic ML predictors to large open‑source language models.

The Machine Learning Enablement initiative builds the core infrastructure that turns complex ML workflows into seamless, self‑service platforms for Etsy’s Applied Scientists and Engineers. The Models Training and Serving Platform team builds and operates the core infrastructure that powers model training and serving across Etsy, including Barista, our internal model‑serving platform, along with a range of hosted model sets and open‑source LLMs. We own the path from a trained model to a production‑ready, observable, and scalable serving endpoint — spanning infrastructure, deployment tooling, and reliability. We work on significant, complex challenges that intersect multiple critical teams and systems, where you can make a rewarding impact. We are light on process and heavy on collaboration, working with many partner teams within our org and beyond to improve our leverage. We are a platform team with a product‑driven mentality, driving innovation using our Machine Learning systems in effective and creative ways.

Responsibilities

• Write high‑quality, production‑grade Python code (additional languages a plus); participate in code reviews and pair programming; contribute to and help drive architectural decisions.
• Build, operate, and improve infrastructure for training and serving ML models on GCP and Kubernetes, with a focus on scalability, reliability, and observability.
• Design and maintain serving infrastructure for open‑source and internally hosted LLMs, including model deployment, resource management, and performance tuning.
• Apply working knowledge of ML fundamentals — including neural network deep learning, latest transformer architectures, and prediction and inference systems — to make sound infrastructure and design decisions.
• Partner cross‑functionally with Applied Scientists to understand model training and serving needs, and translate that understanding into infrastructure that removes friction from their workflows.
• Contribute to system design discussions, weighing tradeoffs across performance, cost, and reliability for infrastructure serving 100M+ users.
• Thoughtfully use generative AI and other productivity tools to work