Musabbir Ahmed Arrafi

Musabbir Ahmed Arrafi

Roboticist || Explorer || Problem Solver


I graduated from the University of Dhaka, Bangladesh with a Bachelors of Science in Robotics and Mechatronics Engineering

My research interest lies at the intersection of computer vision, machine learning, deep reinforcement learning and control systems. Moreover, I would like to see their applications in solving real life problems with field robotics.

I am currently employed as a research assistant at Department of Robotics and Mechatronics Engineering, University of Dhaka. In my free time I try to solve people’s coding problem on StackOverFlow.

  • Computer vision
  • Deep learning
  • Reinforcement learning
  • Remote Sensing
  • Control System
  • Robot vision and perception
  • BSc in Robotics and Mechatronics Engineering, 2018-2022

    University of Dhaka, Bangladesh


Research Assistant
June 2022 – Present University of Dhaka, Bangladesh

Precision Agriculture and Food Safety (Funded by the Ministry of Science & Technology, Bangladesh)

  • Developing an automated plant disease detection pipeline for native crops of Bangladesh utilizing computer vision and deep learning algorithms
  • Working on developing an autonomous shelf-life monitoring system for fruits and vegetables by estimating the weight, volume, moisture content and deterioration state by computer vision algorithms
Research Assistant
October 2022 – Present University of Dhaka, Bangladesh

Landslide susceptibility mapping and flood vulnerability analysis

  • Developing an automated landslide mapping and flood vulnerability analysis pipeline for Bangladesh using data collected from satelite and UAV by deep learning and computer vision
Machine Learning Engineer
April 2023 – October 2023 Dhaka, Bangladesh
  • Developed and deployed an intelligent appointment booking system on GCP with NLP models
  • This project was in collaboration with Dr. Javed Mostafa of University of North Carolina, Chapel Hill
Engineer level 1 (IoT & AI)
January 2021 – February 2023 Dhaka, Bangladesh
  • Contributed to firmware development and automation scripts with C/C++ and python for embedded devices and contributed to backend codebase for IoT products
  • Contributed to the implementation of computer vision projects

Letter of recommendation from DMA

Research Assistant
Supervisor: Dr. Shugata Ahmed
January 2022 – January 2023 University of Dhaka, Bangladesh

Autonomous robot navigation (Final year project)

  • Explored applications of deep reinforcement learning algorithms for autonomous map-less robot navigation in unstructured and unknown terrain by self-supervised method
  • We build a mobile robot that learns to navigate through unseen mapless environments utilizing D3QN and TD3 algorithm
Mentor (Robotics)
March 2019 – December 2022 Dhaka, Bangladesh
  • Mentored gold-medal winning teams of 2020 and 2021 in International Robot Olympiad from Bangladesh
  • Conducted workshops on basic Robotics, ROS1, CAD, Programming in Python
Research Assistant
March 2020 – June 2021 University of Dhaka, Bangladesh

AI driven Autism Detection

  • Benchmarked several models such as SVM, LSTM, VGG16, Inception on publicly available datasets for autism detection through gaze detection, gait Pattern and repetitive behavior monitoring

Skills/ Hands-on experience


  • Python
  • C/C++ for embedded systems
  • Git and GitHub
  • BASH, Linux
  • ROS1, ROS2, Gazebo, OpenAI Gym

Cloud platforms

  • Amazon Web services
  • Google Cloud Platform


  • Machine Learning (Regression, Classification, Clustering, Dimensionality Reduction, Word Embeddings, NLP, Reinforcement learning)
  • Deep Learning (Transfer Learning, CNN, GCNN, NeRF, DRL, vision-Transformers)

Development boards

  • Arduino (AVR and SAMD boards)
  • ESP32, ESP8266
  • Raspberry-Pi
  • Nvidia Jetson Nano, Nvidia Jetson Orin

CAD Tools

  • Fusion 360


  • 3D printing
  • Laser cutting
  • CNC milling
  • Soldering, Welding

Honors & Awards

StackOverFlow - Top 0.09% (73rd) in the world (Aug,2023)
My current stats on StackOverFlow
ROBI datathon 2.0 - Top 10 finalist team
Benchmarked ensemble learning, bagging and boosting models for feature selection and provided a optimal solution based on geospatial industry dataset(approx 40GB) and proposed our solution. Our team ranked within top 10 in the final.
You can find our team(TINY_RME)’s rank in pre-assesment round on Kaggle.
CAD competiton - First Runner Up
Made 3D CAD model and performed motion study and fluid simulation using SOLIDWORKS
Distinguised mentor
Recognized as a distinguised mentor at Bangladesh Robot Olympiad