Portfolio

Michael Hirschmugl

Embedded Systems Engineer working across firmware, hardware, and system integration — with a research background in hardware-efficient ML (quantization & CNN acceleration).

C / C++ ARM / STM32 Linux VHDL / FPGA RISC-V Edge AI & HW/SW Co-design
LinkedIn GitHub IEEE Xplore

I like building practical systems: from real-time firmware to hardware prototypes — and occasionally retro-console hardware.

Selected Projects

open-hw-cnn FPGA / CNN

Open-source hardware accelerator framework for convolutional neural networks on FPGAs.

pipelined-dot-engine VHDL / SV

Pipelined dot-product engine (DOT) built around the reusable MAC core. Separate branches for VHDL and SystemVerilog implementations.

SNES-Drone Hardware

A creative hardware project turning the Super Nintendo's sound chip (SPC700) into a drone synthesizer using a custom cartridge.

These projects highlight my interest in practical HW/SW co-design — from reusable compute blocks to complete embedded hardware builds.

Publications

Fast and resource-efficient CNNs for Radar Interference Mitigation on Embedded Hardware IEEE

Lead author. Hardware-efficient CNN inference for automotive radar interference mitigation, analyzing latency, memory, and power trade-offs across CPU, fixed-point firmware, and FPGA acceleration.

Design and implementation of a calibration solution for automotive RADAR sensors IEEE

Calibration solution design for automotive radar sensors.

Standardized and flexible hardware platform for testing automotive calibration devices IEEE

Hardware platform enabling standardized testing of calibration devices.

Full publication list is available on IEEE Xplore.