Ehsan Kabir

PhD Candidate, Computer System Design Lab, Computer Engineering, University of Arkansas, Fayetteville, USA

Email: ekabir@uark.edu

Phone: +1-479-404-0796

Web: Kabir-Ehsan.github.io

About Me

As a PhD candidate in Computer Engineering at the University of Arkansas, I have been working on accelerating deep learning algorithms using FPGA and GPU platforms. With expertise in designing and implementing neural network accelerators, I’ve worked on various machine learning models, including LSTM, CNN, and MLP, optimizing them for high-rate dynamic applications. My current research focuses on developing a transformer neural network overlay for FPGAs, pushing the boundaries of computational efficiency in AI. I bring over five years of professional experience across multiple roles, including research assistant, teaching assistant, and assistant engineer. My technical expertise spans programming languages like C, C++, Verilog, SystemVerilog, and Python, alongside proficiency with tools like Xilinx Vivado, Cadence and Quartus. I've had the opportunity to work on large-scale engineering projects, from designing low-power digital systems to securing digital circuits against fault injection attacks. I am driven by the challenge of solving complex engineering problems and am eager to contribute to innovative solutions in hardware acceleration, embedded systems, and AI. My work has been published in leading conferences such as FPL, FPT, and FCCM, and I aim to continue advancing cutting-edge technologies through collaborative research and development. I completed my Bachelor of Science in Electrical and Electronic Engineering from Bangladesh University of Engineering and Technology in March 2016. After graduation, I worked as a Lab Instructor at North South University, Dhaka, Bangladesh for a year. Then I joined in Bangladesh Power Development Board as an Assistant Engineer, where he served for two years. I finished my Master of Science degree in Computer Engineering from University of Arkansas, Fayetteville, USA in July 2023. I am expecting to finish my PhD in Computer Engineering from the same university by the end of December 2024.

Interests: Field Programmable Gate Array, Hardware Software Co-design, Machine Learning, Data Science.

Technical Skills:

  • Programming languages: C, C++, VHDL, Verilog RTL, Python, Java, Embedded C, SQL, HTML, CSS.
  • Tools: Xilinx Vivado, Cadence, Matlab, Lumerical, Latex, Arduino, Eclipse, VS code, Anaconda, Jupyter, Git, Proteus, Pspice, Quartus, Openscad, Linux OS.
  • Hardware: FPGA, MicroBlaze & ARM core processor, LabView-National Instruments, Arduino.

Education

University of Arkansas, Fayetteville, USA August 2020 - December 2024

Doctor of Philosophy in Computer Engineering CGPA: 4.00/4.00

University of Arkansas, Fayetteville, USA August 2020 - July 2023

Master of Science in Computer Engineering CGPA: 4.00/4.00

Bangladesh University of Engineering and Technology, Dhaka, Bangladesh February 2011 - March 2016

Bachelor of Science in Electrical and Electronic Engineering

Professional Experience

Research Assistant, University of Arkansas, Computer Engineering August 2020 - December 2024

  • Implementing machine learning algorithms on FPGA.
    Current project: State estimation from high-rate vibration data using LSTM and Transformers on FPGA.

Teaching Assistant, University of Arkansas, Computer Engineering January 2021 – May 2021

  • Graded programming assignments of GPU Programming course.

Teaching Assistant, University of Arkansas, Physics Dept. August 2019 – June 2020

  • Graded course papers, exams and assignments of University Physics 1 & 2 course.
  • Proctored in the lab and course exams.

Assistant Engineer, Bangladesh Power Development Board, Sunamganj, Bangladesh October 2017 – July 2019

  • Ensured uninterrupted electricity supply to 30k consumers by monitoring the operation and maintenance of 33/11kv
    substations, transformers, and other electrical equipment.
  • Supervised a team of 30 sub assistant engineers, electricians, and helpers to run the maintenance operation, solve
    power interruption related issues, and achieve fiscal targets for minimum system loss and maximum bill collection by
    identifying illegal and irregular electric consumption by taking necessary steps to resolve them.
  • Surveyed the area under the substation for new connections and extension of the existing electrical lines to
    distribute power to new customers.

Lab Officer, North South University, Dhaka, Bangladesh September 2016 – July 2017

  • Designed lab manuals to connect theoretical coursework with practical implementation on DC Circuits, Analog
    Electronics, Signal processing, and Control System.
  • Conducted lab classes, experiments, and exams for up to 40 undergraduates at a time.

Research Projects

  • Designed an LSTM neural network using C and RTL language for high-rate dynamic system models on FPGA.
  • Developed an LSTM neural network model for high-rate dynamic systems on Python.
  • Designed a run time programmable convolution and multilayer perceptron neural network accelerator on FPGA.
  • Designed a multilayer perceptron neural network accelerator for high-rate time series signal forecasting on FPGA.
  • Designed a surface plasmon resonance biosensor to analyze lipid molecules and detect diseases like breast cancer, anemia, and diabetes, and increased the sensitivity of the sensor.

Engineering Projects

  • Secured Digital System: Designed a fault injection attack and mitigation technique for a circuit running DES algorithm on FPGA using Verilog RTL and C.
  • Low Power Digital System: Designed a multilayer perceptron neural network on FPGA using Xilinx Vivado Tools and optimized it for low power consumption, high speed, and low resource utilization using high-level synthesis.
  • Recommendation System: Used deep neural network to improve the prediction accuracy of movie ratings from Netflix Datasets using Python on the GPU.
  • Handwritten Digit Classifier: Designed machine learning models such as random forest, decision tree, multilayer perceptron, and convolution neural network for MNIST digit classification using PyTorch and ScikitLearn tools in Python.
  • Smart Car Parking System: Implemented the logic of finding the nearest vacant spot for parking from two different entrances of a parking lot using digital logic gates.
  • n-Bit Cascade-able Comparator Design: Designed a 4-bit cascade-able comparator that can also be used as an n-bit comparator after cascading in CADENCE software.
  • Automated Fire Extinguisher Robot: Built an obstacle detector robot containing a water sprayer using an Arduino Uno Board, fire sensors, light-detecting resistance, and a servo motor.

Publications

  • Ehsan Kabir, Daniel Coble, Joud N. Satme, Austin RJ Downey, Jason D. Bakos, David Andrews, and Miaoqing Huang. "Accelerating LSTM-based High-Rate Dynamic System Models." (Accepted in Field Programmable Logic Conference 2023). Paper Link
  • MD Arafat Kabir, Ehsan Kabir, Joshua Hollis, Eli Levy-Mackay, Atiyehsadat Panahi, Jason Bakos, Miaoqing Huang, and David Andrews. "FPGA Processor In Memory Architectures (PIMs): Overlay or Overhaul?" (Accepted in Field Programmable Logic Conference 2023). Paper Link
  • Coble, Daniel, Joud Satme, Ehsan Kabir, Austin RJ Downey, Jason Bakos, David Andrews, Miaoqing Huang, Adrine Moura, and Jacob Dodson. "Towards online structural state-estimation with sub-millisecond latency." (Accepted in Shock and Vibration Exchange Symposium 2023). Paper Link
  • Ehsan Kabir, Arpan Poudel, Zeyad Aklah, Miaoqing Huang, and David Andrews. "A Runtime Programmable Accelerator for Convolutional and Multilayer Perceptron Neural Networks on FPGA." In International Symposium on Applied Reconfigurable Computing, pp. 32-46. Cham: Springer Nature Switzerland, 2022. Paper Link
  • Atiyehsadat Panahi, Ehsan Kabir, Austin Downey, David Andrews, Miaoqing Huang, and Jason D. Bakos. "High-rate machine learning for forecasting time-series signals." In 2022 IEEE 30th Annual International Symposium on Field-Programmable Custom Computing Machines (FCCM), pp. 1-9. IEEE, 2022. Paper Link
  • Syed Mohammad Ashab Uddin, Sayeed Shafayet Chowdhury, and Ehsan Kabir. "Numerical analysis of a highly sensitive surface plasmon resonance sensor for SARS-CoV-2 detection." Plasmonics 16, no. 6 (2021): 2025-2037. Paper Link
  • Ehsan Kabir, Syed Mohammad Ashab Uddin, and Sayeed Shafayet Chowdhury. "Optimization of surface plasmon resonance biosensor for analysis of lipid molecules." In 2020 2nd International Conference on Advanced Information and Communication Technology (ICAICT), pp. 59-64. IEEE, 2020. Paper Link
  • Syed Mohammad Ashab Uddin, Sayeed Shafayet Chowdhury, and Ehsan Kabir. "A theoretical model for determination of optimum metal thickness in kretschmann configuration-based surface plasmon resonance biosensors." In 2017 International Conference on Electrical, Computer and Communication Engineering (ECCE), pp. 651-654. IEEE, 2017. Paper Link
  • Sayeed Shafayet Chowdhury, Syed Mohammad Ashab Uddin, Ehsan Kabir, and AM Mahmud Chowdhury. "Detection of dna mutation, urinary diseases and blood diseases using surface plasmon resonance biosensors based on kretschmann configuration." In 2017 International Conference on Electrical, Computer and Communication Engineering (ECCE), pp. 662-665. IEEE, 2017. Paper Link