Vishal M

Gained structured industry exposure to end-to-end ML workflows, including, model selection, evaluation metrics, cross-validation, and deployment pipelines Annotated Oriented Bounding Boxes (OBB) for column segmentation in large-scale document images using Xanylabeling, building structured labeled datasets for a document information-extraction pipeline Developed and benchmarked YOLO-based column detection models; built geometric correction modules (un-warping & de-skewing) using OpenCV, increasing OCR-ready column yield by ∼35%; generated ground-truth image pairs for training supervision

Data Science & Machine Learning Intern at Hi-Tech Digital Solutions LLP (2026-01 – 2026-04)

• Gained structured industry exposure to end-to-end ML workflows, including, model selection, evaluation metrics, cross-validation, and deployment pipelines
• Annotated Oriented Bounding Boxes (OBB) for column segmentation in large-scale document images using Xanylabeling, building structured labeled datasets for a document information-extraction pipeline
• Developed and benchmarked YOLO-based column detection models; built geometric correction modules (un-warping & de-skewing) using OpenCV, increasing OCR-ready column yield by ∼35%; generated ground-truth image pairs for training supervision
• Built NLP pipelines for text classification, sentiment analysis, and named entity recognition using NLTK, spaCy, and Hugging Face Transformers
• Performed data preprocessing, feature engineering, and hyperparameter tuning, using Optuna, on real-world client datasets; maintained version-controlled codebases using Git in a collaborative team environment

Integrated M.Sc. in Mathematics – Sardar Vallabhbhai National Institute of Technology, Surat (2021 – 2026)