MatX's mission is to make the world’s best AI models run as efficiently as allowed by physics, bringing the world years ahead in AI quality and availability. We are developing vertically integrated full-stack solutions from silicon to systems including hardware and software to train and run the largest ML workloads for AGI. We are looking for people who are excited about systems-focused ML research.
Responsibilities include:
- Build performance models and tooling to validate and guide scheduling decisions for current and future ML models.
- Write production-grade libraries for efficient distributed training and serving.
- Collaborate with architects, hardware, and software teams to drive solutions from model to metal.
Requirements:
- Bachelor's or Master's degree in Computer Science, Engineering, or a related field
- Strong programming skills in Python
- Expertise in ML frameworks such as JAX, PyTorch, or Tensorflow.
- Deep knowledge of the Transformer architecture.
- Experience with distributed computing, high performance networking, or large-scale ML systems,
Preferred Skills:
Any of the following:
- Hands on experience with flash attention, quantization, pruning, or other systems performance optimizations.
- Fluency in parallelism strategies that balance computation, communication, and memory to improve throughput and latency of large models.
- Experience with performance analysis tools and profilers for large scale systems.
- Solid understanding of computer architecture and low-level optimization techniquesA track record of impactful ML Systems Research, through publication and/or industrial practice.
- Experiences in pre-silicon exploration, post-silicon bringup, and fleet debug.
Compensation: The US base salary for this full-time position is $120,000 - $400,000 + equity + benefits
As part of our dedication to the diversity of our team and our focus on creating an inviting and inclusive work experience, MatX is committed to a policy of Equal Employment Opportunity and will not discriminate against an applicant or employee on the basis of race, color, religion, creed, national origin or ancestry, sex, gender, gender identity, gender expression, sexual orientation, age, physical or mental disability, medical condition, marital/domestic partner status, military and veteran status, genetic information or any other legally recognized protected basis under federal, state or local laws, regulations or ordinances.
All candidates must be authorized to work in the United States and work from our offices in Mountain View Tuesdays-Thursdays.
This position requires access to information that is subject to U.S. export controls. This offer of employment is contingent upon the applicants capacity to perform job functions in compliance with U.S. export control laws without obtaining a license from U.S. export control authorities.
Top Skills
What We Do
MatX designs hardware tailored for the world’s best AI models: We dedicate every transistor to maximizing performance for large models. For these models, we deliver 10× more computing power, enabling AI labs to make models an order of magnitude smarter and more useful. Our hardware would make it possible to train GPT-4 and run ChatGPT, but on the budget of a small startup.
A world with more widely available intelligence is a happier and more prosperous world—picture people of all socioeconomic levels having access to an AI staff of specialist MDs, tutors, coaches, advisors, and assistants.
.png)
.png)
.png)
