Artificial Intelligence Engineer

Radley James; San Francisco (CA); On-site; USD 180,000 - 240,000; Full time

Job summary

Radley James is partnering with a well-backed, early-stage AI company to hire a Member of Technical Staff. This role blends frontier research (≈30%) with serious engineering (≈70%), building environments, datasets and production-ready solutions.

You will tackle model reasoning, RL, alignment, and large-scale evaluations, while shaping training data, synthetic data, and agent systems across distributed infra.

Qualifications

• Top-tier research credentials with publications in leading ML conferences.
• Ability to translate research into production-ready systems.
• Strong background in reinforcement learning, data, and evaluation.

Responsibilities

• Work across model reasoning and post-training to improve performance.
• Engage in RL, alignment and reward design, and large-scale evaluations.
• Develop training data and synthetic data pipelines, and deploy environments.
• Build agent systems and distributed execution to production.

Skills

Research excellence

Production engineering

Python

Distributed systems

RL basics

Education

PhD in ML/CS

Tools

PyTorch

TensorFlow

Python

Job description

We are working with an exceptionally well-backed, early-stage AI company tackling one of the biggest constraints facing frontier models: the availability of high-quality training data.

They’re building the infrastructure and research capabilities needed to push models beyond traditional supervised fine-tuning, across reinforcement learning, synthetic data, simulation, post-training and learning from experience.

They’re now looking for a Member of Technical Staff to join a small, elite team working directly with some of the world’s leading AI research organisations.

This is a rare combination of frontier research and serious engineering. Roughly 30% of the role is research and 70% is building, identifying where current models fail, developing new approaches, building the environments and datasets to address those failures, and getting the work into production quickly.

You’ll work across:

• Model reasoning and post-training
• RL, alignment and reward design
• Large-scale evaluations and benchmarking
• Training data and synthetic data
• Agent systems and distributed execution

The ideal profile combines top-tier research credentials, ideally a PhD with publications at NeurIPS, ICML or ICLR, with a proven ability to turn research into working production systems.

You’ll have genuine autonomy over the problems you tackle and operate much closer to founding-team scope than a conventional research role.