Machine Learning Engineer

Goliath Partners Inc.; San Francisco (CA); On-site; USD 180,000 - 240,000; Full time

Job summary

Goliath Partners Inc. is seeking an ML Infrastructure Engineer in the San Francisco Bay Area to build and scale the infrastructure for training large ML models. You will develop distributed training systems, pipelines, and tooling to accelerate research and production deployments.

You will own GPU load balancing, checkpointing, and experiment management while improving throughput and reliability in a fast-moving environment with strong technical ownership. Bay Area in-person work is expected.

Qualifications

• Strong Python and software engineering fundamentals.
• Experience building ML infrastructure, training infrastructure, or large-scale distributed systems.
• Experience working with PyTorch and modern ML training stacks.
• Strong understanding of distributed computing and GPU-based workloads.
• Experience building reliable systems that support ML research or production ML.
• Comfortable working in a fast-moving environment with significant technical ownership.

Responsibilities

• Build and scale infrastructure for training large machine learning models.
• Develop distributed training systems and improve training efficiency, reliability, and throughput.
• Build data pipelines and infrastructure supporting large-scale ML workloads.
• Improve GPU utilization, compute orchestration, checkpointing, and experiment management.
• Develop tooling that enables researchers and ML engineers to iterate faster.
• Diagnose performance bottlenecks across training, data, and compute systems.
• Help take ML systems from experimentation through production and real-world deployment.

Skills

Python

Distributed computing

ML training stacks

GPU workloads

Production ML

Research support

Tools

PyTorch

Job description

I'm working with a rapidly growing, well-funded AI company building some of the most technically ambitious real-world AI systems today.

They're looking for an ML Infrastructure Engineer to join a highly technical team responsible for building the infrastructure that powers large-scale model training, experimentation, and deployment.

This is a hands-on engineering role for someone who enjoys solving difficult systems problems at the intersection of machine learning, distributed systems, and infrastructure.

What you’ll work on:

• Build and scale infrastructure for training large machine learning models
• Develop distributed training systems and improve training efficiency, reliability, and throughput
• Build data pipelines and infrastructure supporting large-scale ML workloads
• Improve GPU utilization, compute orchestration, checkpointing, and experiment management
• Develop tooling that enables researchers and ML engineers to iterate faster
• Diagnose performance bottlenecks across training, data, and compute systems
• Help take ML systems from experimentation through production and real-world deployment

What they’re looking for:

• Strong Python and software engineering fundamentals
• Experience building ML infrastructure, training infrastructure, or large-scale distributed systems
• Experience working with PyTorch and modern ML training stacks
• Strong understanding of distributed computing and GPU-based workloads
• Experience building reliable systems that support ML research or production ML
• Comfortable working in a fast-moving environment with significant technical ownership

Especially interesting backgrounds include:

• Distributed training and large-scale model training
• ML platforms / internal training infrastructure
• GPU infrastructure and compute orchestration
• Large-scale data infrastructure and pipelines
• Performance optimization for ML workloads
• Infrastructure supporting robotics, embodied AI, computer vision, or other real-world ML systems
• Experience taking ML systems beyond research and into production or physical-world environments

This is a great opportunity for an engineer who wants to work on hard ML systems problems at scale while being much closer to the models and real-world applications than you would be on a traditional infrastructure team.

Bay Area - in person