Hardware Design Engineer

Full time Employment

AI arithmetic Specialty

Apex Compute is at the forefront of revolutionizing AI compute. Our mission is to redefine AI compute architecture by designing efficient, innovative, and cutting-edge hardware and software solutions. We are a passionate team of engineers pushing the boundaries of what's possible in AI hardware, and we offer company stock options to ensure our team shares in our long-term success.

We are seeking a highly motivated Hardware Design Engineer (Floating-Point Architecture) to join our team on a full-time basis. This role is based in Los Altos, CA. As a Hardware Design Engineer, you will play a key role in designing arithmetic units and numerical compute pipelines for our next-generation AI compute hardware. You will work closely with architecture, compiler, and systems teams to implement highly efficient arithmetic datapaths optimized for modern AI workloads.

This includes designing and optimizing floating-point and low-precision arithmetic units supporting formats such as BF16, FP32, FP8, FP4, and INT4, along with specialized hardware for block quantization and efficient numerical processing.

• Design and implement high-performance arithmetic units including adders, multipliers, fused datapaths, and reduction pipelines.
• Develop floating-point and low-precision datapaths supporting BF16, FP32, FP8, FP4, and INT4 arithmetic.

• Design hardware mechanisms for block quantization, data scaling, swamping handling, and subnormal number processing.
• Implement efficient rounding, normalization, and exception handling mechanisms for floating-point pipelines.
• Implement LUT-based and polynomial-based function approximations for operations such as exponentials, reciprocals, normalization, and activation functions.
• Optimize approximation accuracy, latency, and resource utilization.
• Design and optimize arithmetic datapaths for high throughput and efficient hardware utilization.
• Collaborate with architecture and compiler teams to ensure arithmetic units align with system-level performance goals.
• Develop test benches and validation infrastructure to verify arithmetic correctness and numerical stability.
• Work with verification and systems teams to debug and optimize hardware implementations.

• Bachelor's or Master's degree in Electrical Engineering, Computer Engineering, or a related field.

• Strong experience designing floating-point and fixed-point arithmetic units.
• Deep understanding of floating-point formats and numerical behavior, including BF16, FP32, FP8, FP4, and INT
  4.

• Proficiency in Verilog, System Verilog, or VHDL for hardware design.
• Experience implementing adders, multipliers, reduction trees, and numerical datapaths.
• Familiarity with LUT-based or polynomial function approximations.

• Strong understanding of numerical precision, rounding behavior, and subnormal number handling.
• Ability to analyze and optimize arithmetic pipelines for performance, accuracy, and hardware efficiency.
• Experience designing arithmetic units for AI accelerators or high-performance compute architectures.
• Familiarity with quantization techniques such as block quantization or low-precision training and inference.
• Experience with hardware/software co-design or performance-critical compute systems.

You'll have the opportunity to work on revolutionary AI hardware alongside a talented, passionate, and ambitious team. This role provides a chance to grow your technical skills, gain valuable hands-on experience, and contribute to groundbreaking innovations in AI compute hardware. Additionally, as part of our team, you'll be eligible for company stock options, allowing you to share in our long-term success.