Staff ML Platform Engineer for Scalable Graph & MLOps

Fdetimes; United States; Remote; USD 230,000 - 322,000; Full time

Benefits offered by this job

Restricted stock units (RSUs)

Commission opportunities

Medical, dental, and vision insurance

401(k) with employer match

Generous time off

Job summary

Reddit is seeking a Staff ML Infrastructure Engineer to lead development of a platform for large-scale ML models. You will design end-to-end MLOps patterns, support a graph ML codebase, and collaborate with engineers to optimize training time and GPU costs in a distributed environment.

You will also optimize data processing in data warehouses and architect pipelines for massive graph structures, with a strong focus on scalability, reliability, and user-friendly tooling for ML teams.

Qualifications

• 8+ years in ML infrastructure, including model training and deployments.
• Hands-on experience with ML optimization, including memory and GPU profiling.
• Cloud-based ML platform experience (GCP BigQuery, Cloud Storage, IaC).
• Experience administering and integrating MLOps tools for experiment tracking, model serving, and registries (MLflow, Wandb).
• Proficiency with ML languages/frameworks (Python, PyTorch, TensorFlow).
• Experience with distributed training frameworks (Ray, Kubernetes).
• Strong focus on scalability, reliability, performance, and user-centric design.

Responsibilities

• Design end-to-end model lifecycle patterns (MLOps) to boost velocity of development.
• Zero-to-one development and support of a graph ML codebase and platform for scalability.
• Collaborate on performance tuning to improve training time and GPU costs.
• Optimize batch data processing in data warehouses using Beam, Spark, Ray Data.
• Architect pipelines to manage graphs with billions of nodes and edges.

Skills

ML infrastructure

GPU profiling

Python

PyTorch

TensorFlow

Ray

Kubernetes

Terraform

MLflow

Wandb

Tools

GCP BigQuery

Google Cloud Storage

Job description

Reddit is seeking a Staff ML Infrastructure Engineer to lead development of a platform for large-scale ML models. You will design end-to-end MLOps patterns, support a graph ML codebase, and collaborate with engineers to optimize training time and GPU costs in a distributed environment.

You will also optimize data processing in data warehouses and architect pipelines for massive graph structures, with a strong focus on scalability, reliability, and user-friendly tooling for ML teams.