Principal Architect, Performance Analysis and Modeling

At d-Matrix, we are focused on unleashing the potential of generative AI to power the transformation of technology. We are at the forefront of software and hardware innovation, pushing the boundaries of what is possible. Our culture is one of respect and collaboration.

We value humility and believe in direct communication. Our team is inclusive, and our differing perspectives allow for better solutions. We are seeking individuals passionate about tackling challenges and are driven by execution. Ready to come find your playground? Together, we can help shape the endless possibilities of AI.

Working onsite at our Santa Clara, CA headquarters 3 days per week hybrid. Will consider remote in the United States.

The role: Principal Architect, Performance Analysis and Modeling

d-Matrix is seeking outstanding computer architects to help accelerate AI application performance at the intersection of both hardware and software, with particular focus on emerging hardware technologies (such as DIMC, D2D, 3D-DRAM, etc.) and emerging workloads (such as generative inference, etc.). Our acceleration philosophy cuts through the system, ranging from efficient tensor cores, storage, and data movements along with co-design of dataflow and collective communication techniques.

What you will do:

• As a member of the architecture team, you will analyze the latest ML workloads (multi-modal LLMs, CoT reasoning models, video/audio generation)

• You will contribute hardware and software features that power the next generation of inference accelerators in datacenters

• This role requires keeping up with the latest research in ML architecture and algorithms and collaborating with different partner teams, including Product, Hardware Design, Compiler, Inference Server, and Kernels

• Your day-to-day work will include (1) analyzing the properties of emerging machine learning algorithms and workloads and identifying functional, performance implications (2) Creating analytical models to project performance on current and future generations of d-matrix hardware (3) proposing new HW/SW features to enable or accelerate these algorithms

What you will bring:

Minimum:

• BSEE with 10+ years of industry experience or MSEE preferred with 8+ years of industry experience

• Solid grasp through academic or industry experience in multiple of the relevant areas, computer architecture, hardware-software codesign, performance modeling, ML fundamentals (particularly DNNs)

• Programming fluency in C/C++ or Python

• Experience with developing analytical performance models, architecture simulators for performance analysis

• Research background with publication record in top-tier architecture, or machine learning venues is a huge plus (such as ISCA, MICRO, ASPLOS, HPCA, DAC, MLSys etc.)

• Self-motivated team player with strong sense of collaboration and initiative

Equal Opportunity Employment Policy

d-Matrix is proud to be an equal opportunity workplace and affirmative action employer. We’re committed to fostering an inclusive environment where everyone feels welcomed and empowered to do their best work. We hire the best talent for our teams, regardless of race, religion, color, age, disability, sex, gender identity, sexual orientation, ancestry, genetic information, marital status, national origin, political affiliation, or veteran status. Our focus is on hiring teammates with humble expertise, kindness, dedication and a willingness to embrace challenges and learn together every day.

d-Matrix does not accept resumes or candidate submissions from external agencies. We appreciate the interest and effort of recruitment firms, but we kindly request that individual interested in opportunities with d-Matrix apply directly through our official channels. This approach allows us to streamline our hiring processes and maintain a consistent and fair evaluation of al applicants. Thank you for your understanding and cooperation.

Verified from the employer's public Ashby job collection. The provider-supplied title, description, location and application URL were normalized without inferring unstated facts.