Senior ML Research Engineer: Model Alignment & Evaluation

IC Resources; Greater London; Presencial; GBP 90.000 - 160.000; Jornada completa

Descripción de la vacante

Intellectual Capital Resources Ltd in London, United Kingdom, seeks a Senior ML Research Engineer/Scientist to architect and own model evaluation frameworks and post-training pipelines. You will lead a senior technical group to shape the roadmap and research direction for adaptive intelligence systems.

The role requires 5+ years of production ML/LLM experience, deep knowledge of model post-training, LLM evaluation, and reinforcement learning from feedback.

Formación

• Hands-on mastery in model post-training, LLM evaluation, preference data curation, reinforcement learning from feedback, or distillation.
• Strong proficiency in Python and modern deep learning frameworks (PyTorch or JAX).
• Experience leading technical vision and mentoring engineers or researchers.

Responsabilidades

• Design and execute post-training workflows—including preference optimisation, reward modelling, reinforcement learning, and distillation.
• Build rigorous evaluation suites to assess model outputs across contexts and histories.
• Architect strategies for curating high-signal training data using human feedback and synthetic trajectories.
• Evaluate and improve context retention, memory, and adaptation over extended interactions.
• Investigate failure modes and improve inference or retrieval conditioning.

Conocimientos

Python

Experiment design

LLM evaluation

Reinforcement learning from feedback

Mentoring

Herramientas

PyTorch

JAX

Experiment tracking

Descripción del empleo

Intellectual Capital Resources Ltd in London, United Kingdom, seeks a Senior ML Research Engineer/Scientist to architect and own model evaluation frameworks and post-training pipelines. You will lead a senior technical group to shape the roadmap and research direction for adaptive intelligence systems.

The role requires 5+ years of production ML/LLM experience, deep knowledge of model post-training, LLM evaluation, and reinforcement learning from feedback.