Senior ML Engineer — Self-Improving Agent Research

CoreWeave; Bellevue (WA); On-site; USD 182,000 - 242,000; Full time

Benefits offered by this job

Medical, dental, and vision insurance

401(k) with generous match

Paid parental leave

Competitive equity awards

Job summary

CoreWeave is accelerating AI at scale with a focus on reliable, self-improving agents. You will contribute to the CRAFT team's efforts to enable continuous learning in production and validate research ideas on real customer tasks using substantial GPU resources.

The role emphasizes fast delivery, learning on the job, and ownership of impactful ML research and engineering work across CUDA kernels to high-performance LLM tracing dashboards.

Qualifications

• BS/MS/PhD in CS/ML/Robotics or related field required.
• 4+ years ML experience or PhD with 2+ years relevant experience.
• Strong Python skills and PyTorch/JAX experience.
• Deep knowledge of LLM post-training techniques.
• Experience deploying ML models in production environments.
• Sharp researcher with strong problem-solving skills.

Responsibilities

• Generate and validate research ideas to enable continuous learning in production.
• Collaborate with the CRAFT team to test directions on real tasks.
• Contribute to GPU-heavy pipelines and ML tracing dashboards.
• Ship quickly while learning and iterating in a fast-paced setting.

Skills

Educational degree in CS/ML/Robotics

4+ years ML experience

Python & ML frameworks (PyTorch/JAX)

LLM post-training techniques

Production ML deployment

Research & problem solving

Education

Bachelor’s, Master’s, or PhD in Computer Science, ML, Robotics or related field

Tools

Kubernetes

Megatron

Temporal

Postgres

FastAPI

Job description

CoreWeave is accelerating AI at scale with a focus on reliable, self-improving agents. You will contribute to the CRAFT team's efforts to enable continuous learning in production and validate research ideas on real customer tasks using substantial GPU resources.

The role emphasizes fast delivery, learning on the job, and ownership of impactful ML research and engineering work across CUDA kernels to high-performance LLM tracing dashboards.