I am Hirok Roy Rahul, a final-year Computer Science and Engineering student at BRAC University, Dhaka, building at the intersection of AI research, full-stack engineering, machine vision, and automation.
I enjoy turning ideas into working systems, whether that means training ML pipelines, building full-stack AI products, designing automation workflows, or experimenting with computer vision for real-world robotics.
Current focus:
├── AI / ML Engineering
├── Full-Stack AI Products
├── Multimodal Deepfake Detection
├── Computer Vision for AUV Robotics
├── Agentic AI and Automation
└── Research-to-Product Thinking- 🧠 ExplainFake — multimodal deepfake detection with explainable audio-visual reasoning
- ✂️ InstructPrune — adaptive visual token pruning for Vision-Language Models
- 🛍️ Tryora — AI-powered virtual try-on platform with 3D/product intelligence
- 📊 AI-Powered Telecom Churn Prevention — churn-risk ranking and retention analytics
- 🌊 BRACU DUBURI AUV — machine vision for underwater object detection and navigation
- 🤖 Korviora — AI automation ideas for chatbots, e-commerce, and customer support systems
Python · PyTorch · TensorFlow · Scikit-Learn · OpenCV · Pandas · NumPy · spaCy · LangChain · LangGraph
TypeScript · JavaScript · Next.js · React · React Native · Node.js · Express · FastAPI · Tailwind CSS
MongoDB · MySQL · PostgreSQL · Supabase · Prisma · Docker · Git · GitHub · Azure · AWS · Figma
C · C++ · Bash · Linux · Arduino · Embedded Systems · Cisco Packet Tracer
| Project | What it does | Stack |
|---|---|---|
| Tryora | AI-powered virtual try-on platform for e-commerce with image enhancement and 3D model generation | Next.js, Node.js, BullMQ, Prisma, Claid AI, Tripo3D |
| AI-Powered Telecom Churn Prevention | ML system that identifies high-risk telecom customers and supports targeted retention actions | Python, ML, EDA, LightGBM, CatBoost |
| FynmanAI | AI-powered EdTech project built for hackathons and learning support | Next.js, Node.js, OpenAI, WhisperAI, Prisma |
| World Cup 2026 Prediction Pipeline | Predicts FIFA World Cup 2026 fixtures using ML models and Monte Carlo simulation | Python, Scikit-Learn, Pandas |
| ExplainFake | Undergraduate thesis on explainable multimodal deepfake detection | Computer Vision, Audio AI, XAI |
| InstructPrune | Adaptive visual token pruning for efficient Vision-Language Model inference | PyTorch, Hugging Face, VLMs |
- 🥇 Finalist — Infinity AI BuildFest 2026
- 🥇 Finalist — Outskill AI Builders Hackathon
- 🤖 Participant — Band of Agents Hackathon
- ⚡ Enrolled — AMD Developer Hackathon: ACT II
- 🧩 Kaggle — ARC Prize 2026 Paper Track
- ⚽ Kaggle — Soccer Feature Engineering Hackathon
- 🛰️ Kaggle — Hyperspectral Object Tracking Challenge 2026
I like building from zero.
I like research that becomes usable.
I like systems that are clean, practical, and explainable.Outside code and research, I enjoy playing guitar, traveling, and cooking.
``` ::: 


