Hello, I'm
Aspiring researcher and prospective doctoral candidate developing AI solutions for real-world applications.
I am Pranit Das, an aspiring researcher and prospective doctoral candidate with a strong foundation in machine learning and information retrieval. Currently pursuing my M.Tech in Computer Science (Data Science and AI) at NIT Durgapur with a CGPA of 9.37, I am passionate about developing AI solutions that address real-world challenges.
My research focuses on AI for Education, Natural Language Processing, and developing statistical models for data-driven research. I've gained practical industry experience as a Data Engineer at MathCo, where I built scalable data pipelines and visualization solutions that improved team productivity by 20%.
I believe in leveraging AI to create innovative, ethical solutions that make meaningful impacts across education, healthcare, and agriculture. My work combines rigorous research methodologies with practical implementation skills.
Year of Experience
Publications
Projects
Supervised and unsupervised learning, generative modeling and evaluation methodologies
Query understanding and retrieval-augmented systems for enhanced information access
Text classification, semantic similarity, text categorization and language model applications
Developing statistical and ML-based solutions
Designed a framework for personalized learning by integrating context understanding, topic extraction, and Q/A module. Developed novel classification framework to classify user understanding with generative AI models.
Created hierarchical classification model to classify human movement into seven categories. Assembled GY-91 sensor and ESP32-S3 to collect activity data, achieving 92% accuracy on edge devices.
Evaluated the impact of quantum image preprocessing by comparing classification performance with normal inputs. Leveraged transfer learning on pre-trained TensorFlow models trained on ImageNet.
Collaborated in a 4-person team to execute deep learning model for classifying indoor sounds. Quantized and pruned the model (with 5% accuracy loss) for efficient deployment on Raspberry Pi.
17th IEEE CICN, 2025
Developed a context-aware framework with role-based reasoning to classify questions into Bloom's Taxonomy levels. Demonstrated that multi-prompt ensembling with role-based reasoning improves question classification robustness compared to single-prompt baselines.
7th IEEE ICRCICN, 2025
Evaluated ML, DL and Prompt-based generative methods for Bloom's Taxonomy classification across multi-dataset. Identified major generalization failures in supervised learning approaches due to OOV verbs and action verb overlaps.
Jan 2026 - Present
Conducting research in AI for Education, developing methods to reduce learning gaps caused by generic AI-generated content. Assisting PhD researcher in performance benchmarking of image classification methodologies and implementing sound classification models for edge devices.
Aug 2025 - Dec 2025
Served as Lab In-Charge for C Programming Laboratory. Assisted course instructor in preparing lecture resources for the Data Mining course.
May 2025 - July 2025
Worked under faculty supervision on YouTube Video Topic Modeling and Multi-modal Dataset Creation. Implemented topic modeling on video titles to extract high-quality keywords for dataset creation pipeline.
June 2022 - Oct 2023
Built and maintained sales dashboard on Tableau for retail media team, improving performance and productivity by 20%. Designed ETL pipeline architecture and developed front-end tool for GCP cluster recommendations.
Jan 2022 - June 2022
Performed customer segmentation to identify profitable customers, generating leads for credit card sales.
Durgapur, West Bengal, India