Available for research collaborations

Rahul Drabit Chowdhury

Rahul Drabit Chowdhury

AI Engineer & Researcher at BUET. Building at the intersection of evolutionary algorithms, computer vision, and generative AI. Google Summer of Code 2025 contributor.

Get In Touch GitHub LinkedIn
5+
OSS PRs Merged
7+
Research Projects
200+
Students Trained
01

About

I'm a Master's student in Computer Science at BUET, working at the frontier of artificial intelligence. My work spans evolutionary algorithms, deep learning, and generative AI, with a focus on solving real-world scientific and engineering problems.

As a Google Summer of Code 2025 contributor with MBDyn, I implemented a Genetic Algorithm module in C++ for evolutionary optimization. Previously, I worked as an AI Engineer Intern at Technyx, building computer vision and GenAI pipelines.

I believe in open science — I contribute to major projects like OpenVINO, Keras, and AIOpsLab, and co-founded a digital training centre that taught 200+ students during the COVID-19 pandemic.

M.Sc. Computer Science & Engineering
Bangladesh University of Engineering & Technology (BUET)
Dhaka, Bangladesh  ·  2024 – 2026
B.Sc. Mathematics
Mawlana Bhashani Science & Technology University
Tangail, Bangladesh  ·  2018 – 2023  ·  CGPA 3.26/4.00
PGD, Information & Communication Technology
Bangladesh University of Engineering & Technology
Dhaka, Bangladesh  ·  2023 – 2024
02

Experience

Jun – Aug 2025
Remote
Google Summer of Code 2025 Contributor
MBDyn · Genetic Algorithm Module

Designed and implemented a Genetic Algorithm module for evolutionary optimization within the MBDyn multibody dynamics framework. Contributed production-grade C++ code integrating evolutionary algorithm strategies.

C++ Evolutionary Algorithm Genetic Algorithm Open Source
Dec 2024 – Apr 2025
Dhaka, Bangladesh
AI Engineer Intern
Technyx

Developed AI/ML pipelines for computer vision and generative AI applications. Worked with modern deep learning stacks for real-world product deployments, spanning data science workflows and production-ready model serving.

Computer Vision Gen AI Python Data Science
Nov 2020 – Jan 2022
Cox's Bazar, Bangladesh
Co-founder & Trainer
Digital Computer Training Centre

Co-founded and operated a digital training centre during the COVID-19 pandemic, providing free and low-cost instruction in MS Office, Photoshop, and Illustrator to over 200 students from the local community.

Teaching Adobe Suite MS Office Community Impact
03

Research

DNA Binding Protein Prediction — comparing ML and DL frameworks for bioinformatics classification tasks
Ongoing
Knowledge Graph Reasoning for time-related data using Graph Neural Networks
Ongoing
Fuzzy-Based Automated Genetic Algorithm Optimization on HPO of Physics-Informed Neural Networks (PINN)
Ongoing
Unified Deep Learning Approaches — Fusion-based UNET for medical image segmentation
Ongoing
Autonomous Site Relativity Engineering using Reinforcement-Enabled LLM-based Multiagent System
Ongoing
Protein-Protein Interaction Site Prediction using deep learning approaches
Ongoing
Distributed System for Federated Learning — privacy-preserving decentralized model training
Ongoing

Conference Paper

Microcontroller-Based High-Speed and Secured Data Communication System for National Defense using LASER
National Conference on Physics – 2023
Jahangirnagar University, Savar, Dhaka  ·  Bangladesh Physical Society  ·  9–11 March 2023
Poster Presentation
04

Skills

Research Areas
Computer Vision Generative AI Evolutionary Algorithms Machine Learning Quantum Computing Bioinformatics
Frameworks & Tools
PyTorch TensorFlow FastAPI Django LangChain LangGraph CrewAI Qiskit Docker
Languages
Python C++ C Rust Java JavaScript SQL MATLAB Fortran
05

Open Source

openvinotoolkit/openvino
PR #32554
Implemented translator for PyTorch's cross_entropy_loss operation in OpenVINO's PyTorch frontend (log_softmax + NLL loss with reduction modes and weights) and added tests.
keras-team/keras
PR #22146
Extended Keras to support bitwise_left_shift and bitwise_right_shift operations in the OpenVINO backend, ensuring correct mapping of bitwise operators during inference.
microsoft/AIOpsLab
PR #126
Resolved a bug in Helm integration by correcting parameter passing to helm.status() in Prometheus telemetry, fixing a runtime crash.
openml/openml-python
PR #1679
Refactored tests to replace live network calls with local mocks for deterministic and offline test execution; no production code changes.
openvinotoolkit/openvino
PR #34343
Implemented translator for the quantized::relu6 operator via a clamp + requantization approach and added regression tests for the OpenVINO inference engine.
06

Projects

GAT_ASP-UNet
Ensemble-Based U-Net for Medical Image Segmentation combining graph attention networks, atrous spatial pyramid pooling, and multi-scale feature fusion for robust biomedical image analysis.
PyTorch U-Net GAT Medical Imaging Deep Learning
StackDPP vs TransBind
Comparative study of ML and DL approaches for DNA binding protein prediction — benchmarking traditional stacking classifiers (StackDPP) against a Transformer-based model (TransBind) on bioinformatics datasets.
Python Bioinformatics Transformers ML vs DL
Blockchain EHR System
Decentralized Electronic Health Record system enabling privacy-preserving medical document sharing and storage using blockchain, federated learning, IPFS, and smart contracts.
Go JavaScript Python Blockchain Federated Learning
ML Learning Projects
Collection of machine learning projects completed during self-study from Coursera, ITB, and CUET — covering supervised/unsupervised learning, neural networks, and applied ML algorithms.
Python scikit-learn NumPy Coursera Mar 2024
Restaurant Website
Full-featured order-enabled restaurant management site with table reservations, menu browsing, and order processing — built as a desktop application with a GUI frontend and relational database backend.
Java Java Swing MySQL Mar 2024
Team Graph Analysis Report
Network analysis using graph theory for humanitarian aid team network study — analyzing team structure, relationships, and information flow patterns using advanced graph algorithms and visualization techniques.
Graph Theory Network Analysis Data Visualization Humanitarian
Research Survey: Maximum Edge Colorable Subgraph
Comprehensive survey on the Maximum Edge Colorable Subgraph Problem — exploring theoretical foundations, computational complexity, algorithms, and applications in graph coloring and optimization research.
Graph Coloring Combinatorics Research Survey Optimization
07

Achievements & Certifications

3rd
Innovation Cohort 4, University Innovation Hub Program, BUET — Data Transmission using Laser Light
Aug 2025
4th
BRAC University AI Hackathon — Web-based Farmer AI Assistant (Agriculture Team)
May 2025
5th
Innovation Cohort 2, University Innovation Hub Program, BUET — AI-Enhanced Drone for Rescue & Search
May 2025

Certifications

Specialization
Machine Learning Specialization
Coursera  ·  DeepLearning.AI  ·  Andrew Ng  ·  Credential ID: 4CWOM9EO60MW

Let's build something together

Say Hello.

Open to research collaborations, interesting problems, and conversations about AI/ML, evolutionary computing, or quantum.