Machine Learning Engineer (Training Infrastructure)

Genesis AI

Greater London

On-site

GBP 100,000 - 160,000

Full time

14 hours ago
Be an early applicant

Get a reply from this employer

Your resume and cover letter, tailored to exactly what they’re hiring for.

Get past ATS filters

Job summary

Genesis AI is seeking an engineer to take ownership of the infrastructure powering our distributed training stack. You will design, build and optimize multi-node PyTorch training systems on GPU clusters, focusing on scalability, robustness and high utilization.

You will profile the stack end to end, reduce bottlenecks, and implement efficient low-level components using CUDA, cuDNN and Triton. You’ll collaborate with researchers to translate goals into reliable, high-performance training

Qualifications

  • 5+ years of deep engineering experience in distributed systems, ML infrastructure or high-performance computing.
  • Strong low-level performance skills across CUDA, cuDNN or Triton.
  • Hands-on experience scaling PyTorch training across large GPU clusters.
  • Track record of improving hardware-software interactions with measurable gains in utilisation, throughput or training time.
  • The ability to work closely with researchers and turn goals into reliable training systems.

Responsibilities

  • Design, build and optimise distributed training systems in PyTorch across multi-node GPU clusters.
  • Profile the training stack end to end, identify bottlenecks and reduce time to convergence.
  • Write and integrate efficient low-level implementations using CUDA, cuDNN, Triton and custom kernels where they make a measurable difference.
  • Improve CPU-GPU compute balance, memory management, data throughput and networking for better hardware utilisation.
  • Build monitoring and debugging tools to diagnose performance regressions and distributed system issues.

Skills

Distributed systems
ML infrastructure
High-performance computing
CUDA optimization
PyTorch training

Tools

CUDA
cuDNN
Triton

Job description

Genesis is a full-stack physical AI company building general purpose robots that can operate in any environment, on any task. We develop our models, our general-purpose simulation engine, Genesis World, and our hardware in-house, including our dextrous hand and training glove.

Following our first VLA, GENE 26.5, we're working towards scaling training across thousands of GPUs using millions of hours of multimodal data, alongside large-scale reinforcement learning in simulation. We're looking for an engineer who can take real ownership of the infrastructure that makes this possible.

You'll:
  • Design, build and optimise distributed training systems in PyTorch across multi-node GPU clusters, with a focus on scalability, robustness and high utilisation.
  • Profile the training stack end to end, identify bottlenecks and reduce time to convergence.
  • Write and integrate efficient low-level implementations using CUDA, cuDNN, Triton and custom kernels where they make a measurable difference.
  • Improve CPU-GPU compute balance, memory management, data throughput and networking so workloads make effective use of the hardware.
  • Build monitoring and debugging tools that help the team quickly diagnose performance regressions, failed runs and distributed systems issues.
What you'll bring
  • 5+ years of deep engineering experience in distributed systems, ML infrastructure or high-performance computing.
  • Strong low-level performance skills across CUDA, cuDNN or Triton, including kernel optimisation, CPU-GPU interactions and data movement.
  • Hands-on experience scaling PyTorch training across large GPU clusters using data, context, pipeline and model parallelism.
  • A track record of improving hardware-software interactions, with measurable gains in utilisation, throughput or training time.
  • The ability to work closely with researchers, understand what they're trying to achieve and turn that into reliable, efficient training systems.
Get your free, confidential resume review.

or drag and drop your file here.

Similar jobs

Similar jobs worth comparing

Distributed ML Training Systems Engineer
Distributed ML Training Systems Engineer

Genesis AI • Greater London

On-site
GBP 100,000 - 160,000
Senior AI Training Infra Engineer - Distributed PyTorch
Senior AI Training Infra Engineer - Distributed PyTorch

Genesis AI • Greater London

Hybrid
GBP 120,000 - 170,000
Training / AI Infrastructure
Training / AI Infrastructure

Genesis AI • Greater London

Hybrid
GBP 120,000 - 170,000
Member of Technical Staff, ML Performance
Member of Technical Staff, ML Performance

Odyssey • Greater London

On-site
GBP 70,000 - 90,000
Senior ML Infrastructure Engineer (Research Initiatives) - Systems Integrator
Senior ML Infrastructure Engineer (Research Initiatives) - Systems Integrator

Hamilton Barnes Associates Limited • United Kingdom

Hybrid
GBP 90,000 - 130,000
Significant stock option packages
Remote-first working setup
Fully paid travel and accommodation
+1
Founding Performance Engineer
Founding Performance Engineer

Axioma Search SARL • Greater London

Hybrid
GBP 120,000 - 180,000
Founding Software Architect
Founding Software Architect

Axioma Search SARL • Greater London

Hybrid
GBP 90,000 - 130,000
Visa sponsorship
Relocation assistance
Founding Distributed ML Infra Architect
Founding Distributed ML Infra Architect

Axioma Search SARL • Greater London

Hybrid
GBP 90,000 - 130,000
Visa sponsorship
Relocation assistance
Founding Engineer - ML Systems (up to £250k)
Founding Engineer - ML Systems (up to £250k)

Dex • Greater London

On-site
GBP 90,000 - 150,000
Senior Data Infrastructure Engineer - Petabyte-Scale AI Pipelines
Senior Data Infrastructure Engineer - Petabyte-Scale AI Pipelines

Genesis AI • Greater London

Hybrid
GBP 120,000 - 180,000