NER using BERT
Named Entity Recognition system built using a fine-tuned BERT model. High-quality dataset preprocessing, tokenization, and evaluation with F1 metrics.
"Building intelligent systems across ML, backend, and mobile."
I am a final-year B.Tech CSE student and a hands-on engineer building end-to-end systems that combine machine learning, backend engineering, and Android applications. I enjoy taking research ideas and turning them into production-ready solutions.
I’ve built projects across NLP (NER with BERT, RAG systems), Computer Vision (pothole detection), and Android apps (Kaksha Connect, Alarm app). I also build secure backend services using Java and JWT for authentication.
A few projects that showcase my work across NLP, CV, backend, and mobile development.
Named Entity Recognition system built using a fine-tuned BERT model. High-quality dataset preprocessing, tokenization, and evaluation with F1 metrics.
End-to-end RAG pipeline: document ingestion, chunking, vector embeddings, FAISS retrieval, and generator integration for contextual Q&A.
Computer Vision model to detect and localize road potholes using a custom dataset and YOLO-style detection model. Includes a simple Android demo for field testing.
Secure authentication service using Java (Spring Boot) with JWT-based stateless auth, refresh tokens, and role-based access control.
A classroom/learning Android app with messaging, assignments, and notifications. Built with Kotlin and Room DB for local storage.
Custom alarm manager app with repeat rules, notifications, and a simple clean UI. Focused on reliability and battery-efficiency.
B.Tech — Computer Science & Engineering
Currently in final year (or specify year). Relevant coursework: Data Structures, Algorithms, Machine Learning, Operating Systems.
Want to collaborate or see the code? Reach out using the links or send an email.