I am a Master's researcher in Transportation Engineering at the Faculty of Engineering, Chulalongkorn University, Thailand, working at the intersection of Artificial Intelligence and Smart Transportation Systems.
My research applies Natural Language Processing (NLP), Explainable AI (XAI), and Large Language Models (LLMs) to uncover public sentiment toward urban mobility — with a current focus on Bangladesh's rapidly evolving multi-modal transport landscape.
mahbub = {
"institution" : "Chulalongkorn University, Bangkok, Thailand",
"degree" : "MSc in Transportation Engineering",
"department" : "Civil Engineering — Transportation Division",
"research" : ["NLP", "XAI", "LLM", "Sentiment Analysis", "Smart Mobility"],
"languages" : ["Python", "R", "MATLAB", "SQL", "LaTeX"],
"currently" : "Building ML + XAI + LLM pipeline for transport sentiment analysis",
}| Domain | Topics |
|---|---|
| 🚇 Smart Transportation | ITS, Urban Mobility, Multi-modal Systems, Developing Countries |
| 🗣️ NLP & Text Mining | Sentiment Analysis, ABSA, Multilingual (Bangla/Banglish/English) |
| 🔍 Explainable AI (XAI) | SHAP, LIME, DiCE Counterfactuals, Feature Importance |
| 🧠 Large Language Models | Open-source LLMs, Research Automation, LLM-in-the-loop |
| 📊 Machine Learning | Classification, Ensemble Methods, Hyperparameter Tuning (Optuna) |
| 🌆 Sustainable Mobility | Green Transport, EV Adoption, CAV, MaaS |
A multilingual NLP · ML · XAI · LLM research pipeline for emerging-market transport discourse
| Dataset | 8,000 synthetic social media records (Facebook · YouTube · Twitter · Instagram · Reddit) |
| Languages | Bangla (40%) · Banglish (35%) · English (25%) |
| Sentiment | Negative · Positive · Neutral (50/30/20 %) |
| Aspects (ABSA) | 12 aspects — fare, safety, comfort, punctuality, overcrowding, cleanliness, infrastructure, accessibility, service quality, route coverage, driver behavior, staff behavior |
| Transport Modes | 13 modes — MRT-6, BRTC, AC Bus, Local Bus, Intercity Train, CNG, Rickshaw, Pathao, Shohoz, Uber, inDrive, Launch/Ferry, Tempo |
| Validation | 12/12 checks ✅ · 99.5% unique texts · Date range: Dec 2022 – Dec 2024 |
| Phase 2 | ML Classification + SHAP/XAI + LLM-based policy insight generation (in progress) |
