👋 I'm Shanin Hossain

an AI Engineer & Research Enthusiast passionate about:

Machine Learning, Deep Learning, and Generative AI
Computer Vision & Natural Language Processing
Healthcare Informatics and Medical Imaging

+8801778198423

Dhaka, Bangladesh

June 15, 2001

Download CV

🚀 Explore my Projects below, or head to the page to interact with me!

Experience and Education

AI Engineer at HawkEyes Digital Monitoring Limited

Jan 2024 – Present

Prepared datasets, trained and fine-tuned models, and optimized accuracy for real-world AI applications.

Research Assistant (Remote) at Queensland University

June 2024 – Present

Developed a hybrid graph-based brain glioma grading system using 3D MRI datasets.

Junior Front-End Developer

Jan 2024 – Feb 2024

Designed a responsive news portal template for publishers and bloggers.

Research Lab Member (HIRL Lab - DIU)

Jan 2023 – Dec 2023

Worked on medical image disease detection, sentiment analysis, and integrated computer vision with advanced ML algorithms.

B.Sc. in Computer Science and Engineering at Daffodil International University (DIU)

Jan 2020 – Dec 2023

Projects

🚀 Unilever Bangladesh AI Pipeline

Built a YOLO-based AI pipeline for object detection, voice recognition, auto-segmentation, facial recognition attendance, and eyeball tracking R&D, deployed with FastAPI on Vast.ai cloud servers

Python Ultralytics OpenCV face-recognition FastAPI Vast.ai

🔍 Cigarette Display Analysis for BAT

Led cigarette display analysis using image processing, creating algorithms for box segmentation, size measurement, distance calculation, and sequence validation, with AR/VR for 3D visualization to ensure compliance.

Warp Perspective Instance Segmentation Sequence Generation Sorting

🧐 AI-Powered Face & Voice Login

Designed a real-time multimodal authentication system using deep face recognition (MTCNN + InceptionResnetV1 + KNN-GNN) and speech-based identity verification (ML, DL & SpeechBrain), achieving secure and efficient biometric login.

KNN-SVM-GNN MFCCS MTCNN InceptionResnetV1 SpeechBrain

📄 OCR Automation

Developed an automated OCR pipeline to extract item names, quantities, and prices from market receipts that demonstrates end-to-end implementation of computer vision and data extraction solutions for retail automation.

Regex OpenCV Tesseract OCR Google API

🖼️ GUI Tool for Image Perspective Transformation

The application enables users to upload or paste images, apply perspective transformations interactively, and save or copy the transformed images. It is designed to be user-friendly, to manipulate images with ease.

Python Tkinter Pillow NumPy

🧬 Graph Radiomics for Breast Tumor Detection

This project applies Gamma correction and CLAHE to enhance ultrasound images, from which radiomics features are extracted. Pretrained models and a CNN-LSTM architecture are utilized alongside a proposed GNN to model feature relationships and accurately classify benign or malignant.

Radiomic Features GCN GAMMA-CLAHE Transfer Learning CNN-LSTM

💬 Portfolio Q&A AI Chatbot

This project is an AI-powered Portfolio Q&A Chatbot built using Flask and LangChain. It uses GPT-4.1-Nano as the LLM and retrieves answers from your PDF portfolio data through a RAG pipeline. The system combines Hugging Face sentence embeddings with Pinecone vector storage.

GPT-4.1-Nano Pinecone Hugging Face Embedding LangChain

📝 Bangla LLM Text Summarizer

Build Bangla text summarizer using Unsloth’s Gemma-3-4b, fine-tuned with LoRA on custom Bangla data, exported to GGUF for lightweight inference, and powered by Ollama for offline use. The sleek Flask + Tailwind CSS chat UI lets users paste articles and get instant, accurate summaries.

Unsloth GGUF Ollama NLP Flask HTML Tailwind