Computer Architect (Generative AI)

TL;DR

Computer Architect (Generative AI): Designing scalable, high-performance compute architectures and connectivity topologies for next-generation AI silicon with an accent on advanced memory subsystems and low-latency interconnects. Focus on driving concept-to-silicon ownership and developing executable hardware specifications to optimize LLM performance.

Location: Must be based in Mountain View, New York City, or San Francisco

Salary: $200,000 – $500,000

Company

A high-performing semiconductor firm founded by industry veterans, pushing the boundaries of silicon design for Generative AI and LLMs.

What you will do

• Drive the design of scalable, high-performance compute architectures and connectivity topologies tailored for modern AI workloads.
• Design advanced memory subsystems, including low-latency, high-throughput interconnects and cache hierarchies.
• Collaborate with hardware, firmware, and software teams to ensure specifications meet cross-functional requirements.
• Take concept-to-silicon ownership for key features at the subsystems and chip level, from specification to production.
• Deliver executable specifications as models to describe hardware architecture.

Requirements

• Proven track record of driving silicon architecture for key features from specification to production.
• Proficiency in SystemVerilog, Python, C/C++, Rust, or Bluespec for chip architecture and silicon design.
• Experience delivering innovative features in chips and systems related to ML, compute, networking, or storage.
• Strong ability to debug architectural issues at the silicon and system level across design, validation, and software teams.

Nice to have

• Experience writing firmware and software code to test architectural features at the product level.