PhD student, currently focusing on building and optimizing FPGA-based inference accelerators for 3D CNNs. I take special interest in Hardware/Software co-design, FPGAs and heterogeneous architectures.
Currently working on FPGA-based 3D CNNs accelerators and hardware-aware optimizations.
Worked on hardware resource estimation for the implementation of Deep Neural Networks(DNN) on FPGAs using machine learning algorithms.
Developed applications for mixed signal devices (ADCs) from electrical schematic, PCB layout design to dedicated firmware and software. The main two projects I worked on are: the MCP3564 ADC Evaluation Board and the MCP3564 Weight Scale Demo.
Developed a Medical Precision Thermograph which is based on precision thermistors with 0.05°C accuracy. The device is dedicated for measuring body temperature in order to provide additional information a doctor can use to detect medical conditions that otherwise would be detected through X-ray imaging. I designed the thermistor measurement circuit and simulated it in LTSpice, created the schematic design of the whole device in KiCAD and worked on the PCB layout design. I also programmed the embedded software in C and tested the device using an environmental chamber which could be programmed to maintain constant temperatures for a desired amount of time.
Diploma Thesis: TBD
Diploma Thesis: FPGA-based balancing of an Inverted Pendulum
Check the thesis here .
Diploma Thesis: Precision Thermograph for Medical Use (supported by FotoNation)
Check the thesis here .
. D. Diaconu, L. Petrica, M. Blott and M. Leeser, "Machine Learning Aided Hardware Resource Estimation for FPGA DNN Implementations," 2022 IEEE International Parallel and Distributed Processing Symposium Workshops (IPDPSW), Lyon, France, 2022, pp. 77-83, doi: 10.1109/IPDPSW55747.2022.00022