北

北京知存科技有限公司 Tech Stack

Semiconductor design and AI inference optimization for embedded systems

Semiconductors 北京, 北京 11–50 employees

北京知存科技 is an engineering-focused semiconductor company built around chip simulation, hardware design optimization, and embedded AI inference. The tech stack reveals a dual focus: classical EDA tooling (Synopsys, Cadence) paired with modern AI frameworks (PyTorch, Megatron-LM, NCNN, TensorFlow Lite) and simulation platforms (QEMU, Gem5, SystemC). Active projects span SoC functional modeling, in-memory computing noise optimization, and LVGL display engine adaptation—indicating work on low-power inference and edge deployment.

Tech Stack 43 technologies

Core StackC++ Python PyTorch Linux QEMU Gem5 SystemC RISC-V ARM C/C++ LVGL C GCC UART Ethernet I2C Synopsys Cadence CUDA Megatron-LM NCNN GPU FPGA TensorFlow Lite SNPE Bash RTOS Verilog Unix Perl+13 more

What 北京知存科技有限公司 Is Building

◆Challenges

Reducing computational error
Digital backend design flow optimization
Yield improvement
Production abnormal handling
Low power consumption
High performance
Algorithm complexity
Reducing software hardware resource consumption

▲Active Projects

Digital backend design process optimization and maintenance
Chip hardware interface optimization
Ai inference framework development
Subsystem verification
Software simulation platform development and integration
Soc chip functional and performance modeling
Lvgl display engine adaptation
Chip error modeling
Noise optimization method development
Noise optimization for in-memory computing chips

Hiring Activity

Minimal35 roles · 0 in 30d

Department

Engineering

33

Data

2

HR

1

Ops

1

Seniority

Senior

18

Mid

14

Director

2

Lead

2

Manager

1

Company intelligence

Find more companies like 北京知存科技有限公司 by tech stack, pain points and active projects

Get started free

About 北京知存科技有限公司

北京知存科技 designs semiconductors with emphasis on chip simulation, hardware-software co-optimization, and embedded AI inference. The company operates from Beijing with a lean, engineering-heavy team (33 engineers across senior and mid levels) working on SoC modeling, digital backend design flows, and AI inference framework development. Their project list suggests focus on performance and power efficiency for edge devices, including work on memory-computing chips and display subsystems. Active challenges center on yield improvement, computational error reduction, and low-power execution.

Headquarters北京, 北京

Company Size11–50 employees

Hiring MarketsChina

Frequently Asked Questions

What is 北京知存科技's tech stack?

Simulation: QEMU, Gem5, SystemC. Design: Synopsys, Cadence, Verilog. AI: PyTorch, Megatron-LM, NCNN, TensorFlow Lite, CUDA. Cores: C++, Python, RISC-V, ARM, Linux, RTOS.

What is 北京知存科技 working on?

SoC functional and performance modeling, AI inference framework development, digital backend design optimization, chip error and noise modeling, LVGL display engine adaptation, and subsystem verification.