Senior ML Systems Engineer: RL Training & GPU Optimisation

Axiōma Search; New York (NY); On-site; USD 150,000 - 210,000; Full time

Job summary

Axiōma Search is seeking a hands-on ML researcher to improve models after initial training, with a focus on reinforcement learning and post-training methods. You’ll work on scalable experiments and the systems needed to run them efficiently at scale.

You’ll span learning algorithms and infrastructure, moving from RL rollouts to GPU profiling, identifying bottlenecks and delivering throughput and memory improvements across distributed training.

Qualifications

• Hands-on Python programming with strong problem-solving skills.
• Hands-on experience with PyTorch and model training.
• Understanding of reinforcement learning or LLM post-training.
• Experience with distributed training and GPU systems.
• Strong experimental judgement and ability to work across ML research and systems engineering.

Responsibilities

• Build and improve supervised fine-tuning, preference optimisation and RL methods.
• Work with approaches including PPO, GRPO and SDPO.
• Own training loops from rollout generation through to policy updates and checkpointing.
• Improve training throughput, GPU utilisation and memory efficiency.
• Profile and fix bottlenecks across distributed training.
• Investigate instability and differences between training and inference.
• Use real model failures to improve rewards, training data and overall performance.

Skills

Python

Reinforcement learning

Model training

GPU systems

Experimentation

ML research and systems engineering

Tools

PyTorch

Distributed training

CUDA

Ray

Job description

Axiōma Search is seeking a hands-on ML researcher to improve models after initial training, with a focus on reinforcement learning and post-training methods. You’ll work on scalable experiments and the systems needed to run them efficiently at scale.

You’ll span learning algorithms and infrastructure, moving from RL rollouts to GPU profiling, identifying bottlenecks and delivering throughput and memory improvements across distributed training.