International research award
Best Paper Award
2026 IEEE 9th International Conference on Industrial Cyber-Physical Systems
Co-first author · Paper · Award report
Resource-constrained embedded systems · Energy-aware computing
Hanze Li
I study and build complete embedded systems across power electronics, custom PCBs, firmware, FPGA/Linux, computer vision, and on-device intelligence.
Computer Science undergraduate at ShanghaiTech University, undergraduate researcher at METAL Lab, and President of GeekPie Association.
PEAK power-management board, independently designed and assembled
Selected recognition
National CPU design competition
Second Prize
2026 National Student Computer System Capability Competition, CPU Design Track
Team captain · System details · Official result
Shanghai electronic design competition
First Prize
2026 TI Cup Shanghai Undergraduate Electronic Design Competition
Team captain · System details · Official result
Featured research and systems
My work connects power availability, hardware, embedded software, and on-device intelligence. PEAK is the clearest example of that approach.
PEAK | Self-Powered TinyML Sensor Node
A battery-free embedded sensing platform combining custom power hardware with on-device inference.
I independently developed the prototype from schematic and PCB design through embedded firmware, TinyML deployment, wireless reporting, and evaluation.
- 172 nA
- Power-board quiescent current
- Sole developer
- Hardware through TinyML
- Custom PCB
- Designed and assembled
Read the project
Manuscript in preparation
Competition systems
Team results are shown separately from the subsystem work I personally led.
National second prize
MIKU LA32R Out-of-Order SoC
My responsibility: team captain; CPU performance and timing, FPGA system integration, and Linux bring-up.
Four-issue SpinalHDL CPU on Artix-7, closed at 100 MHz with a 4.831× speedup over the competition baseline, plus Linux 5.14, X11, USB HID, and networking.
Shanghai first prize
TI Cup Ball-Balancing Vehicle
My responsibility: team captain; K230 vision system and mechanical integration.
Calibrated 88.9 FPS ball tracking with approximately 1 mm static error, integrated with STM32F407 control and the vehicle-mounted mechanism.
Additional engineering work
Selected hardware and software systems beyond the projects above.
24-Channel Dual-Spectrum Line Sensor
Designed the schematic, dense PCB layout, SMT assembly, enclosure, and MCU firmware. Cascaded-latch multiplexing provides 1 kHz quasi-synchronous infrared and visible-light sampling with ambient-light rejection.
SparkLive
Local multimodal livestream operations software combining Qwen, Whisper/VAD, OBS WebSocket, FastAPI/WebSocket, React, TTS, and digital-human backends.
I captained the team and contributed to system integration and local deployment.
Research
I study systems in which power availability, hardware behavior, scheduling, and algorithms must be designed together.
Energy Management Design Guide for Self-powered Ambient IoT Devices: Considering Energy and Power Ratios Among Source, Storage, and Load
Proceedings of the 9th IEEE Conference on Industrial Cyber-Physical Systems, Perth, Australia, 2026, pp. 1–6. Co-first author.
Developed a two-dimensional energy supply-demand model and translated it into design rules for converter selection, control policy, and storage sizing.
Best Paper Award at the IEEE International Conference on Industrial Cyber-Physical Systems.
Anti-Poaching Patrol Prioritization
Carnegie Mellon University · WWFProf. Fei Fang
Independently developing a spatiotemporal decision-support system that supports wildlife-protection patrol prioritization.
Engineering scope
Circuit and PCB systems
Power-path design, analog signal conditioning, schematic and PCB layout, SMT assembly, board bring-up, and measurement.
Demonstrated in PEAK and DUALiNEEmbedded and real-time software
C/C++, nRF52/nRF54, STM32, ESP32, K230, Zephyr, BLE, UART protocols, Wi-Fi, and MQTT.
Demonstrated in PEAK and TI CupFPGA and Linux integration
Scala/SpinalHDL, Verilog, LA32R, RISC-V, Vivado, timing closure, Linux, drivers, and Buildroot.
Demonstrated in MIKUOn-device intelligence
TFLM, INT8 deployment, feature extraction, CanMV vision, dataset design, measurement, and evaluation.
Demonstrated in PEAK and TI Cup
President of GeekPie Association
September 2025 to present
I lead ShanghaiTech's 66-member student technology association and direct its embedded systems and intelligent hardware program. My work includes technical direction, budgets, external collaboration, member development, laboratory procedures, and engineering events.
During the 2025-26 academic year, GeekPie reported more than 20 technical events and 780 event attendances across workshops, open-source infrastructure, project training, and engineering competitions.
- 66
- members
- 20+
- annual events
- 780+
- event attendances
Research collaboration and embedded systems opportunities
Shanghai, China · Email is the best way to reach me.
MaoTechMail [at] gmail [dot] com