Tech Stack

Description

You will architect and build the next-generation ML data platform at Roblox, including a scalable Feature Store and training data platform, enabling AI applications across recommendations, search, safety, and more. Your work supports billions of daily interactions and impacts millions of users.

Requirements

• 5+ years experience in AI/ML data infrastructure, feature stores, or training data infrastructure
• Experience with NoSQL databases like DynamoDB, Cassandra, CockroachDB
• Expertise in building scalable ML data pipelines for batch and real-time environments
• Strong software engineering skills with distributed computing frameworks (Spark, Flink, Ray) and real-time data systems (Kafka, Redis, DynamoDB)
• Deep knowledge of ML Feature Stores (Feast, Tecton, Vertex AI Feature Store) and embedding/vector search

Responsibilities

• Develop a scalable Feature Store supporting batch and real-time feature computation and serving
• Build Roblox's first training data platform with point-in-time joins and feature injection
• Engineer large-scale ML data pipelines using streaming and batch ETL for billions of records per day
• Build a feature engineering framework to empower ML engineers to author and productionize millions of signals
• Collaborate with cross-functional AI teams on next-gen AI infrastructure