LLM Training & Continual Learning Research Engineer

Collision AI; San Francisco (CA); On-site; USD 120,000 - 180,000; Full time

Job summary

Collision AI in San Francisco is looking for a researcher to develop training methods and infrastructure for frontier cybersecurity models, focusing on post-training and continual learning from real security work.

You’ll own the full experimental cycle—from hypothesis to deployed models—identifying signals in production telemetry, building datasets, testing algorithms, and evaluating model performance, while collaborating with founders to shape the product and research direction.

Qualifications

• Experience training large language models.
• Proficiency in PyTorch, JAX, or a comparable ML framework, with the engineering skills to turn research ideas into working systems.
• A master’s degree or relevant research experience.

Responsibilities

• Develop training methods and infrastructure behind frontier cybersecurity models.
• Own the full experimental cycle—from hypothesis to deployed models.
• Collaborate with founders to shape the product direction and research roadmap.

Skills

LLM training

Experimental judgment

Distributed training

Open-source contributions

Independent thinking

Education

Master's degree or relevant research experience

Tools

PyTorch

JAX

Job description

Collision AI in San Francisco is looking for a researcher to develop training methods and infrastructure for frontier cybersecurity models, focusing on post-training and continual learning from real security work.

You’ll own the full experimental cycle—from hypothesis to deployed models—identifying signals in production telemetry, building datasets, testing algorithms, and evaluating model performance, while collaborating with founders to shape the product and research direction.