SEHR SHAMIR

I’m a Computer Science undergrad with hands-on experience building scalable backend systems and applying ML/DL/NLP to real-world datasets. I enjoy turning messy data into reliable pipelines, models, and APIs that can actually run in production.

On the ML/NLP side, I’ve built an end-to-end NLP pipeline for Twitter user classification using topic modeling (LDA) and contextual embeddings (BERT), plus text preprocessing and clustering evaluation. I also developed a hate speech detection pipeline for Hinglish datasets using TensorFlow/Keras, tokenization, embeddings, and model evaluation with precision, recall, and F1-score.

On the backend side, I’ve developed RESTful APIs using Node.js, Express.js, and MySQL, implemented JWT authentication and RBAC for Admin/Customer roles, and integrated Google Maps API for real-time tracking and route optimization. I’ve also automated CI/CD pipelines using GitHub Actions and improved delivery workflows using Dijkstra’s algorithm.

I’m motivated by building systems that combine strong data engineering with practical modeling—optimized through hyperparameter tuning, cross-validation, and thoughtful validation strategies. I’m also actively sharpening my fundamentals through Data Structures & Algorithms work (LeetCode) and applied research in NLP-based user profiling.