Description

NextSilicon is reimagining high-performance computing. Our accelerated compute solutions leverage intelligent adaptive algorithms to vastly accelerate supercomputers, driving them forward into a new generation. Our new software-defined hardware architecture enables HPC & AI to fulfill their promise of breakthroughs in all fields of advanced research. 

At NextSilicon, everything we do is guided by three core values:

• Professionalism: We get exceptional results through professionalism and unwavering dedication to quality and performance.
• Unity: Collaboration is key to success; we foster a work environment where every employee feels valued and heard.
• Impact: We're passionate about developing technologies that make a meaningful impact on industries, communities, and individuals worldwide.

The AI software team is building solutions that enable the latest AI models and workflows to run efficiently on on NextSilicon’s unique hardware architecture—unlocking a new level of performance

We are seeking a talented engineer with extensive experience in modelling, analyzing and optimizing AI workloads and end-to-end flows. This position lies at the intersection of computer science, hardware architecture and ML computational algorithm

Requirements

• B.Sc. degree in software engineering, computer science, or a related field. M.Sc. or Ph.D. is a plus
• Proficiency in Python and C/C++ 
• 5+ years of experience in software engineering
• 2+ years hands-on experience with profiling and analyzing AI models
• Excellent command of PyTorch or TensorFlow
• Expertise in competitive performance analysis is a strong advantage
• At least one of the following:
• Demonstrated contribution to relevant open-source projects
• 
  
• Significant experience with ML accelerators, high-performance computing, and GPU or CPU architectures

Responsibilities

• Model AI workload behavior on custom hardware and software
• Identify bottlenecks and propose optimizations in hardware architecture and software.
• Define next generation software requirements and specifications