Senior Machine Learning Engineer

You will improve model capability through training, post-training, and rigorous experimentation. Working with applied AI engineers, environment engineers, and domain experts, you will turn useful data and feedback into measurable model improvements. You will own the path from training data and experiment design through reproducible checkpoints and evaluation results. This is a senior, hands-on individual-contributor role., * Build supervised fine-tuning, preference optimization, and reinforcement learning pipelines for language models and AI agents, choosing methods that fit the problem.

• Prepare, validate, and version training datasets with documented provenance and clear separation of training, validation, and held-out evaluation data.
• Establish baselines and controlled experiments that distinguish useful capability gains from data leakage, benchmark overfitting, and measurement noise.
• Work with domain experts and environment engineers to develop feedback, preference data, reward signals, and tasks that reflect intended applications.
• Diagnose training instability, regressions, memory constraints, and throughput bottlenecks; collaborate with infrastructure engineers to use GPU resources effectively.
• Deliver documented checkpoints, reproducible configurations, evaluation findings, and known limitations; verify improvements in representative workflows with applied AI engineers.

• Strong Python and PyTorch skills, with practical experience training or fine-tuning substantial neural networks and explaining your technical choices.
• Understanding of optimization, model architectures, learning objectives, and the factors that affect training stability and generalization.
• Experience with reproducible ML workflows, dataset versioning, experiment tracking, checkpoint management, and reliable evaluation.
• Ability to debug across data, model, and compute layers, including practical experience with GPU workloads and distributed training.
• Evidence of independently delivering meaningful ML work, communicating results clearly, and changing your approach when experimental evidence warrants it.

Preferred experience

• Experience with language-model post-training, preference learning, reinforcement learning, reward modeling, or agents that interact with tools and environments.
• Experience using expert feedback or scientific data, and evaluating failure patterns, uncertainty, and performance across relevant task groups.