X

Xcelerium Tech Stack

Fabless semiconductor processor design for edge AI and wireless communications

Computer Hardware Manufacturing Irvine, CA 51–200 employees Founded 2020 Privately Held

Xcelerium designs domain-specific processors for edge workloads combining wireless, radar, vision, and AI—a narrow but hardware-intensive focus reflected in their stack: Synopsys, RISC-V, CUDA, TensorFlow, and MLIR compiler work. The company is adopting MLIR, a signal they're building custom compiler infrastructure rather than relying on off-the-shelf toolchains. Early-stage hiring (6 open roles, all engineering, mostly senior) and recent focus on signoff automation and physical design suggest they're moving from architecture into production tape-out.

Tech Stack 32 technologies

Core StackC++ Python MATLAB TensorFlow PyTorch NumPy Pandas Synopsys Simulink GNU Radio RISC-V CUDA OpenCL TensorFlow Lite TinyML OpenGL C/C++ Assembly Raspberry Pi BeagleBone Arduino ARM Cortex-M Verilog SystemVerilog UVM Assembly Language IREE TensorRT XLA Keras+1 more

AdoptingMLIR

What Xcelerium Is Building

◆Challenges

Reducing time-to-closure
Optimizing ppa efficiency

▲Active Projects

Software framework for business and product goals
Domain-specific dialects for compute kernels
Modern physical design flow implementation
Automation of signoff processes
Performance-tuned transformation pipelines
Mlir compiler passes

Hiring Activity

Minimal6 roles · 0 in 30d

Department

Engineering

6

Seniority

Senior

5

Mid

1

Company intelligence

Find more companies like Xcelerium by tech stack, pain points and active projects

Get started free

About Xcelerium

Xcelerium is a fabless semiconductor company founded in 2020 by veterans from Qualcomm, Broadcom, and Intel. The company develops high-performance, domain-specific processors optimized for converged edge workloads that combine wireless communications, radar, image processing, and machine learning. Rather than building general-purpose chips, they target narrow application domains where specialized hardware delivers significant performance or power advantages. The organization is US-based with engineering-focused hiring and minimal recent hiring velocity, consistent with a pre-revenue or early customer validation phase.

HeadquartersIrvine, CA

Company Size51–200 employees

Founded2020

Hiring MarketsUnited States

Frequently Asked Questions

What is Xcelerium's tech stack?

Xcelerium uses Synopsys for design, RISC-V and ARM instruction sets, CUDA and OpenCL for compute, TensorFlow/PyTorch/TensorRT for ML workloads, Verilog/SystemVerilog for RTL, and MLIR for compiler infrastructure. MATLAB and Simulink appear in modeling; GNU Radio and real-world platforms (Raspberry Pi, BeagleBone) in validation.