Neuron Collectives Software Engineer (AI)

TL;DR

Neuron Collectives Software Engineer (AI): Developing and optimizing collective communication algorithms and topologies for Trainium AI training chips with an accent on compute and bandwidth utilization. Focus on implementing high-performance C/C++ collective patterns and co-optimizing software with silicon to scale LLM training.

Location: Must be based in Cupertino, CA, USA

Salary: $165,200 - $223,600 annually

Company

Annapurna Labs, part of AWS, develops purpose-built AI training chips (Trainium) and networking solutions for ML and HPC workloads at hirify.global scale.

What you will do

• Enhance collective algorithms and topologies to optimize AI training performance.
• Identify compute and bus bandwidth bottlenecks using Neuron Explorer.
• Develop and optimize C/C++ implementations of collective communication patterns.
• Co-optimize software and Trainium silicon in close collaboration with hardware teams.
• Optimize collective operations to scale AI compute across the data center.
• Build and maintain analysis frameworks and automation solutions for training topologies.

Requirements

• 3+ years of professional software development experience.
• 2+ years of experience in design or architecture of systems (scaling and reliability).
• Proficiency in at least one software programming language.
• Location: Must be based in Cupertino, CA

Nice to have

• 3+ years of experience with the full software development life cycle (SDLC), including code reviews and build processes.
• Bachelor's degree in Computer Science or equivalent.

Culture & Benefits

• Comprehensive health insurance (medical, dental, vision, prescription).
• 401(k) matching and restricted stock units (RSUs) as part of the compensation package.
• Paid time off and parental leave.
• Flexible working hours and a strong organizational commitment to work-life balance.
• Strong mentorship culture with access to principal-level engineers and directors.