As part of our small and growing Machine Learning team, you'll help develop a platform for the end-to-end machine learning lifecycle: rapid prototyping, full-scale training, deploying, monitoring, maintaining models, and iterating on modeling ideas. You'll work on a system which is used for large-scale machine learning. The team's training platform is used for models on all major surface areas (Ads, Home Feed, Recommendations), which score tens of millions of pins every second. Projects you'll work on includes building a lineage-tracking workflow engine for reproducibility of models and creating an unified feature store for teams to share features about Pinterest entities (pins, users, boards).
What you'll do :
- Model management, analysis, monitoring, and deployment tools.
- Efficient prototyping tools for bringing up new ML use cases.
- Build systems for features, training datasets, and models. These systems keep production models running smoothly, by integrating training automation, monitoring, lineage, and more.
- Build prototyping and analysis tools that allow large teams to collaborate and iterate on model quality.
- Work extensively with ML engineers across Pinterest to understand their requirements, pain points, and build them generalized solutions.
What we're looking for :
- At least 4 years of industry experience in software engineering.
- Experience with large-scale production machine learning, and knowledge of at least one ML framework, such as Tensorflow, PyTorch, or equivalent.
- Experience with big data computation frameworks (e.g. Hadoop, Cascading/Scalding, or Spark).
- Strong software design skills.
- Expertise in Java/Scala and Python.