Staff Software Engineer, Machine Learning Platform

Overview

Discord is used by over 200 million people every month primarily for gaming. The Machine Learning Platform team builds, maintains, and scales the tools and services that support the full model development lifecycle across all ML applications.

What You'll Be Doing

As a Staff Software Engineer, Machine Learning Platform at Discord, you will:

• Design and build platforms for ML engineers and data scientists.
• Evaluate and integrate new ML frameworks and tools including LLMs and generative AI.
• Partner with model builders to ship high-quality models quickly.
• Champion best practices in feature engineering and ML infrastructure.
• Build and maintain shared ML foundations including datasets and embedding models.

What You Should Have

Requirements include:

• 8+ years experience as a software or ML engineer with large-scale data and distributed systems.
• 4+ years experience in building platforms/infrastructure for engineering teams.
• 2+ years experience in building machine learning platforms covering orchestration systems.
• Proven ability to establish and execute on clear, actionable technical direction.
• Experience leading technical initiatives across diverse teams.

Bonus Points

• Experience with large-scale production systems on Kubernetes.
• Real-time data processing expertise (Kafka, Spark, or Flink).
• Experience with modern ML frameworks (PyTorch, TensorFlow, or JAX).
• Experience deploying Generative AI applications and evaluation frameworks.
• Experience in managing infrastructure costs for efficiency and scalability.

Why Discord?

Discord plays an important role in the future of gaming. It is a dynamic, multiplatform environment that helps people connect before, during, and after games. Join a mission-driven team and help shape the future of ML-driven user experiences.

Key skills/competency

• Machine Learning
• Platform Development
• Distributed Systems
• Data Infrastructure
• Feature Engineering
• LLM
• Generative AI
• Production Systems
• Kubernetes
• Pipeline Orchestration