Frontier AI Research Engineer — Open-Model Pre-Training

Prime Intellect, Inc.; Oregon; Hybrid; USD 150,000 - 350,000; Full time

Benefits offered by this job

Remote or SF office option

Visa sponsorship

Relocation assistance

Team off-sites & learning events

Job summary

Prime Intellect seeks engineers to advance open frontier AI by taking open models from architecture experiments to full pre-training runs. You will influence model design, training strategies, and system reliability across the training stack.

You will work on scalable ML workflows, including MoE and attention experiments, with an emphasis on reproducibility and performance. Remote or SF-office options offered to suitable candidates.

Qualifications

• Hands-on experience training language models and modifying training code.
• Strong Python skills and experience with PyTorch or a comparable framework, with debugging distributed training.
• Practical understanding of model architectures and how optimizers affect convergence.
• Ability to design ablations to answer research questions and explain results.
• Good judgment on when to investigate independently or escalate to a colleague.

Responsibilities

• Build and extend pre-training code in prime-rl to turn small experiments into full training runs.
• Evaluate architecture changes (MoE, attention) against quality and training cost.
• Profile training jobs and improve end-to-end throughput with hardware-aware execution.
• Ensure long runs are resilient with reliable checkpointing and recovery.
• Investigate training dynamics and numerical stability while preserving convergence.
• Run controlled ablations and keep setups reproducible.
• Collaborate with data and evaluation researchers on training mixtures and long-context experiments.

Skills

Python

Distributed training debugging

Ablation design

Model optimization understanding

Team collaboration

Research decisions

Tools

PyTorch

Profiling tools

Job description

Prime Intellect seeks engineers to advance open frontier AI by taking open models from architecture experiments to full pre-training runs. You will influence model design, training strategies, and system reliability across the training stack.

You will work on scalable ML workflows, including MoE and attention experiments, with an emphasis on reproducibility and performance. Remote or SF-office options offered to suitable candidates.