Performance Engineer, Large-Scale ML Systems

Anthropic; New York (NY); Hybrid; USD 280,000 - 850,000; Full time

Benefits offered by this job

Competitive compensation

Optional equity donation matching

Generous vacation and parental leave

Flexible working hours

Job summary

A technology-focused organization in New York is seeking a Performance Engineer to enhance its large-scale machine learning systems. The role involves identifying and solving systems problems while developing robust solutions. Ideal candidates will have software engineering experience, an interest in machine learning, and a collaborative mindset. As a member of a dedicated team, you will work on innovative projects that shape the future of AI. Competitive compensation and benefits are offered, including flexible working options.

Qualifications

• Experience at supercomputing scale is preferred.
• Strong candidates should have a track record of solving large-scale systems problems.
• Desire to learn more about machine learning research.

Responsibilities

• Identify systems problems in large-scale ML algorithms.
• Develop systems to optimize throughput and robustness.
• Work on projects like low-latency sampling for language models.

Skills

Software engineering experience

Machine learning experience

Results-oriented

Flexibility

Pair programming

Concern for societal impacts

Education

Bachelor's degree in a related field

Tools

GPU/Accelerator programming

ML frameworks

Operating systems internals

Job description

A technology-focused organization in New York is seeking a Performance Engineer to enhance its large-scale machine learning systems. The role involves identifying and solving systems problems while developing robust solutions. Ideal candidates will have software engineering experience, an interest in machine learning, and a collaborative mindset. As a member of a dedicated team, you will work on innovative projects that shape the future of AI. Competitive compensation and benefits are offered, including flexible working options.