About the Role
As an ML Platform Engineer on the ML Platform Engineering team, you will be instrumental in elevating our core ML platform to its next level of performance, reliability, and scalability. You'll work on the critical infrastructure that directly enables all of Afresh's Machine Learning and Applied Science teams to innovate faster and deliver impact. Your contributions will empower our product suite, including our flagship Prediction Engine, to power replenishment decisions on more than 15% of all produce sold in the United States.
What You'll Do
- In your first 3 months, you might deliver a feature that helps generalize model configuration, enables no-code model deploys for our various ML solutions, or vastly improves integration testing across our ML systems.
- By the end of your first 6 months, you will have owned the implementation of significant scalability improvements and additions to our ML platform. This might include new feature pipelines that power our recommendation engine, or work to stand up the first instance of real-time inference at Afresh.
Skills and Experience
- BS in Computer Science or a relevant technical field.
- 3+ years of professional software development experience with a proven track record of shipping high-quality applications and services.
- Experience working collaboratively with machine learning engineers, data scientists, or applied scientists on large-scale software projects involving machine learning models.
- Deep expertise in library design, API design, data structures, and algorithms.
- Strong familiarity with Python.
- Experience working collaboratively with machine learning engineers, data scientists, or applied scientists on large-scale software projects involving machine learning models.
- You possess a genuine curiosity about ML modeling (e.g., demand forecasting, state estimation, ordering policy). You aren't just building "pipes"; you want to understand what is flowing through them.
- You have an understanding of how scientists work and build tools that bridge the gap between a research notebook and production-grade software.
Tech Stack: Our backend is pure Python (NumPy, Pandas, Torch, PySpark, Cython, orchestrated in Airflow). We use Databricks as our data warehouse. While we'd like you to have very good familiarity with Python, many of our problems are stack-agnostic.