Foundational ML Researcher - Learning & Training Dynamics - USA in London

Sentiro Partners is working with a head-turning market leader. We’re building models that are trained under real constraints, not academic ones.

This role sits inside a focused, senior research group focused on how large models learn, not just how they perform on benchmarks. You will have direct ownership over training dynamics, optimization behaviour, and failure modes of large-scale learning systems.

This is not a product ML role. This is not post-hoc evaluation. This is not incremental fine-tuning work.

If your background is in frontier labs, advanced ML research, or deep systems-aware model training, this role is designed to feel familiar, likely more demanding. You have likely published at top conferences, such as NeurIPS, ICML or ICLR.

What you’ll work on:

• Designing and modifying training loops for large-scale models
• Studying learning dynamics, optimization behaviour, and generalization
• Stress-testing models under distribution shift and adversarial conditions
• Building signals that distinguish real progress from benchmark overfitting
• Collaborating with researchers who care about first-principles understanding, not paper counts

What makes this different:

• You own entire research threads, end to end
• You can change how models are trained, not just how they’re evaluated
• Feedback cycles are tight and unforgiving
• Success is defined by robustness and signal quality, not publications

Who this is for:

You might be a fit if you’ve worked on:

• Large-scale model training or optimization
• Reasoning, generalization, or learning dynamics
• Systems-aware ML (where infra and modeling meet)
• Frontier-lab research that felt too abstract or too slow

We are intentionally not optimizing for:

• Narrow domain specialists
• Pure infra engineers
• Product-driven ML roles

Environment:

• Small, senior research team
• Deep technical autonomy
• In-office collaboration (East Coast)
• Compensation and resources competitive with top-tier research labs
• Unlimited compute

Requirements:

• Post-graduate degree in machine learning, computer science, an NLP adjacent discipline, or commercial experience developing novel machine learning algorithms.
• Research depth e.g. demonstrable experience in internships or employment in leading research labs or research environments.

Location: East Coast USA. This role is being advertised across the US. Full visa sponsorship will be provided for international talent.