Software Systems Engineer, Data Labeling Engineering - Early Career

About the Team

Help teach our self-driving vehicles how to see and understand the world.

The Data Labeling Engineering team designs, builds, and operates hybrid human/machine data labeling tools and pipelines that power autonomous vehicle machine learning models within General Motors' AV organization . We operate in the intersection of software engineering , data engineering , and AI/ML , defining the strategies, tooling, and quality controls that create reliable training data at scale. Our tools and platform are used by thousands of users and consumers.

We own a modern full‑stack architecture including TypeScript/React, Python, GraphQL, Golang , and ML model services , which powers data‑annotation pipelines and machine‑led training data solutions at foundation ‑ model scale . We partner closely across AI/ML engineers , Product Operations , Product Management , Data Science , and other ML Platform groups.

About the Role

As an early-career Software Engineer on the Data Labeling Engineering team, you will build tools and services that help machine learning teams create high-quality training data for autonomous driving. Your work may span frontend experiences, backend services, data pipelines, machine learning integrations, and quality systems used by labelers, ML engineers, and operations teams.

This role is designed for a recent college graduate or engineer early in their career who wants to own meaningful pieces of a platform, grow their technical expertise, and work directly on systems that enable the next generation of AV capabilities. You will learn from experienced engineers while contributing to production systems and developing depth across frontend, backend, data, and ML-adjacent technologies.

What You’ll Do

• Level up how ML teams work with data

Develop automation and tooling that give ML engineers deep insight into labeling workflows and data quality (e.g., efficiency dashboards, auto‑QA, autolabel review tools), reducing iteration time from idea to trained model.

…