Lead Developer at a US-based digital agency delivering production React apps driving ~$2K daily revenue. AI Engineering Intern building RAG pipelines and agentic workflows. Published Researcher with Taylor & Francis book chapter on Medical BERT architectures. 3x Dean's List (3.89 CGPA), 36+ GitHub stars.
- CGPA: 3.89/4.00 | 3x Dean's List | Full Academic Scholarship
- 1st Runner-Up, CS Fest 2024 App Showcase (20+ teams)
- Coursework: ML, NLP, Computer Vision, Database Systems, Software Engineering
GPA: 5.00 / 5.00 | Science | Rajshahi Board
Accepted at IEEE QPAIN 2026 (IEEE Xplore, Scopus Indexed). Proposes a clustering-based topic modeling framework for analyzing and decoding research trends from large-scale academic literature.
Benchmarked 5+ Medical BERT architectures (BioBERT, PubMedBERT, ClinicalBERT) on PsyTAR dataset. Achieved SOTA multi-label ADR classification using fine-tuned LLMs with attention mechanism analysis for healthcare NLP.
- Engineered NLP pipeline: 52,702 reviews → 85-90% accuracy with TF-IDF + MultinomialNB
- 17-task ML workflow with 5,000+ features | MIT Licensed
Developed an optimized, cross-platform web scraping system using Python and Scrapy to collect and clean news articles from major Bangladeshi newspapers. Implemented high-performance batch runners, date-range filtering, logging, monitoring, and automated data export to Excel/CSV with a unified database design.
PyTorch, Hugging Face Transformers, BERT, BioBERT, scikit-learn, NLTK, TF-IDF, LLM Fine-tuning
Java, Python, C++, Kotlin, C#, SQL, JavaScript
MySQL, SQLite, SQL Server, JDBC, Connection Pooling, REST APIs, FastAPI, 3NF Database Design
Git, GitHub, Linux, Docker, Jupyter, Neovim, CMake
Authored peer-reviewed book chapter benchmarking 5+ Medical BERT architectures for multi-label ADR classification. Achieved SOTA results on PsyTAR dataset using PyTorch and Hugging Face.
Outperformed 20+ competing teams with Unreal Engine desktop game development. Demonstrated technical proficiency in game design and physics and delivered winning pitch to industry judges.