Quadric
Designing licensable processor IP for on-device AI inference.
Company facts
- Founded
- 2016
- Employees
- 75
- Headquarters
- United States
- Type
- Private
- Funding
- Series C · $30M · 2026
- Subsidiaries
- Quadric Japan
Open positions
Staff Deep Learning Compiler Engineer
Senior engineer designing and optimizing Quadric's deep learning compiler stack (MLIR, TVM, LLVM) that lowers PyTorch/TensorFlow/ONNX models onto its GPNPU edge-AI processor — graph optimizations, quantization, kernel codegen, benchmarking, and hardware-software co-design with the architecture team.
Forward Deployed Engineer
Forward Deployed Engineer at Quadric who owns customer evaluations of its Chimera GPNPU edge-AI chip from first technical conversation to design win — porting workloads, profiling, and writing quantized custom kernels in C++/Python, plus building reusable agentic (Claude Code–based) field-engineering tooling. Remote within the US.
Senior Software Engineer, Compiler (LLVM)
Quadric is redefining edge AI with the industry's first General Purpose Neural Processing Unit (GPNPU), enabling developers to run both neural network inference and conventional C++ code on a single programmable…
Head of Marketing
Quadric is redefining edge AI with the industry's first General Purpose Neural Processing Unit (GPNPU), enabling developers to run both neural network inference and conventional C++ code on a single programmable…
Staff SoC Design Engineer
Staff SoC Design Engineer at Quadric (Pune office) who owns SoC subsystem development, third-party IP integration (NoC, PCIe/UCIe, DDR, security), and PPA/timing-closure work on the company's General Purpose NPU for edge AI. Core tools: SystemVerilog/Verilog, VCS, and Verilog/C co-simulation.
Staff SoC Design Engineer
Staff SoC Design Engineer at Quadric, an edge-AI silicon company in downtown Burlingame, CA. You'll own SoC subsystem development and third-party IP integration (NoC, PCIe/UCIe, DDR, security), drive PPA optimization and timing closure through tapeout, and package designs for customer releases. Core stack: SystemVerilog/Verilog, VCS, Verilog/C co-simulation.
Senior/Staff Deep Learning Compiler Engineer
Designs and optimizes the deep learning compiler stack (MLIR, TVM, LLVM) for Quadric's General-Purpose Neural Processing Unit, building lowering pathways from PyTorch/TensorFlow/ONNX down to optimized kernels and partnering with hardware teams on edge AI performance.
Deep Learning Compiler Intern
An on-site internship in Burlingame, CA working with senior compiler engineers on Quadric's CGC neural network compiler, building and debugging compiler passes (layout selection, memory allocation, code generation) that lower NN IR to C++ for the Chimera GPNPU edge-AI chip. Core tech: Python, C++, IR transformations, and ML compiler frameworks like TVM and MLIR.
Systems Software Engineer
Systems Software Engineer at Quadric building the low-level stack for its General-Purpose Neural Processing Unit (GPNPU) edge-AI chip — owning components like the instruction set simulator, loader/assembler, remote debugger, and C++ host/device APIs. Work is heavily modern C++, done in close co-design with the hardware team.
Regional Sales Executive- North America
Quadric is redefining edge AI with the industry's first General Purpose Neural Processing Unit (GPNPU), enabling developers to run both neural network inference and conventional C++ code on a single programmable…
Senior SoC Design Verification Engineer
Senior SoC Design Verification Engineer at Quadric, an edge-AI processor startup: architect and build a scratch-built SoC-level UVM verification environment for GPNPU-based SoCs, develop directed/random tests, verify 3rd-party IP (NoC, PCIe/UCIe, DDR), and drive coverage using SystemVerilog/UVM, VCS, Verilog/C co-simulation, assembly, and Python.
Senior Hardware Engineer - Micro-Architect
Senior hardware engineer at Quadric (Burlingame, CA) who owns microarchitecture definition and RTL implementation of their General Purpose Neural Processing Unit in SystemC/SystemVerilog, drives PPA optimization, and carries designs through timing closure to tapeout for an edge-AI processor startup.