Machine Learning Infrastructure Engineer

business Meta location_on Seattle, WA (Hybrid) apartment Hybrid Full Time

Compensation

$175K – 250K/yr

bolt Apply for this role

calendar_month

Published Date

3 weeks ago

apartment

Work Arrangement

Hybrid • Seattle, WA

group_add

Open Positions

3 openings

workspace_premium

Experience Level

Senior

overview Role Overview

About the opportunity

Engineer distributed training fabrics and petabyte-scale feature stores supporting global ranking algorithms.

Join Meta's AI Systems team to architect the distributed computing platforms behind PyTorch and internal deep learning clusters. You will solve non-trivial scaling problems related to high-bandwidth interconnects, fault-tolerant checkpointing, and memory-efficient tensor sharding across 100,000+ GPU nodes.

task Core Responsibilities

What you will do

check_circle Develop fault-tolerant training primitives in PyTorch supporting automatic checkpoint save/resume.
check_circle Optimize networking pipelines utilizing RoCEv2 and InfiniBand for ultra-low latency all-reduce communication.
check_circle Profile CPU-GPU data ingestion pipelines to maximize GPU compute utilization percentages.
check_circle Design self-healing cluster orchestration tools for long-running pretraining workflows.

verified_user Candidate Profile

What we are looking for

arrow_circle_right 5+ years experience in systems programming (C++, Python, Rust) and distributed systems.
arrow_circle_right Strong familiarity with PyTorch internals, FSDP, DeepSpeed, and high-performance networking.
arrow_circle_right Experience troubleshooting large-scale compute infrastructure in hyperscale cloud environments.

code_blocks Technologies & Competencies

Skills & Tech Stack

PyTorch C++ Distributed Systems InfiniBand FSDP Python

handshake Direct Referral Network

Why candidate applications stand out

verified

Verified Technical Credentials

Applications include direct proof-of-work repositories and instructor verification endorsements.

person_check

Fast-Track Hiring Visibility

Direct internal referral channels through enterprise partners bypass automated resume discard filters.

lock Confidential Pipeline

Apply for this role

Submit your résumé (PDF, DOC or DOCX up to 5 MB) for direct consideration by the Meta hiring team.

Full name *

Email address *

Phone

Current Location

Upload Résumé *

LinkedIn Profile

GitHub / Portfolio URL

Cover Note

Need 1-on-1 Interview Mentorship?

Work directly with industry practitioners on technical mock interviews, ATS resume positioning, and system design drills.

Explore Career Support arrow_forward