I’m a Computer Science Engineering student focused on Machine Learning, NLP, Computer Vision, and Retrieval Augmented Generation (RAG). I enjoy building practical AI systems that solve real problems and turning concepts into deployable projects.
Over the last few years, I’ve built projects across classical ML, deep learning, computer vision, NLP, and full stack AI applications using Python, TensorFlow, Flask, FastAPI, Docker, PostgreSQL, and modern ML workflows.
- 📘 Published author of a Machine Learning Projects book on Amazon KDP
- 📄 Published research paper: Mood2Mail: A Lightweight Real-Time Email Tone Analyzer using Classical Machine Learning
- 📊 Published datasets on Kaggle with 60+ downloads
- 🧠 Built multiple end-to-end AI/ML projects across NLP, Computer Vision, Recommendation Systems, RAG, and Deep Learning
- ⚡ Experienced with deployment, APIs, Docker, PostgreSQL, FastAPI, Flask, and model integration
- 📈 Completed Kaggle competitions including Titanic and House Prices
- 🔍 Currently exploring advanced RAG systems and MLOps workflows
- Python
- Java
- C++
- SQL
- JavaScript
- HTML/CSS
- Scikit-learn
- TensorFlow
- Pandas
- NumPy
- OpenCV
- Librosa
- NLP
- CNNs
- Recommendation Systems
- Retrieval Augmented Generation (RAG)
- Flask
- FastAPI
- Docker
- PostgreSQL
- REST APIs
- Git & GitHub
An intelligent Retrieval Augmented Generation system that allows users to ask questions from uploaded PDF documents.
- PDF ingestion pipeline
- Semantic retrieval
- Context-aware question answering
- RAG workflow implementation
Python, NLP, Embeddings, Vector Search
A computer vision powered search engine that retrieves visually similar images based on uploaded input images.
- Image similarity search
- Feature extraction using deep learning
- Visual embeddings pipeline
- Computer vision based retrieval system
Python, Computer Vision, Deep Learning
A lightweight NLP-powered application that detects the tone of emails and provides real-time feedback before sending.
- Tone classification using classical ML
- TF-IDF based NLP pipeline
- Real-time prediction system
- Flask backend with custom frontend
- Multiple tone categories including Friendly, Formal, Aggressive, Anxious, and Passive
Python, Scikit-learn, Flask, HTML, CSS, JavaScript
A real-time waste classification system that identifies waste categories using deep learning.
- Image-based waste prediction
- CNN-based image classification
- Real-time inference
- Practical sustainability-focused use case
TensorFlow, CNNs, Python
A deep learning application that recognizes handwritten digits from uploaded images or a drawing board.
- CNN-based handwritten digit recognition
- Interactive drawing board support
- Real-time prediction
TensorFlow, CNNs, Flask
An ML-powered system that predicts the genre of uploaded music files.
- Audio feature extraction using Librosa
- MP3 and WAV support
- ML-based genre prediction
- End-to-end deployment
Python, Librosa, Scikit-learn, Flask
- AI Complaint Triage & Routing System
- BiasX-Ray – Real-Time Bias Detector
- Fake Job Posting Detector
- Movie Recommendation System
- Real-Time Sign Language Detector
- StyleNet – Fashion Image Classifier
- Car Mileage Predictor
- University Time Bank Platform
- Dyslexia AI Assistant
- Published ML research paper
- Published author of a Machine Learning Projects book
- Solved 580+ problems on GeeksForGeeks (101-day streak, rank 159 among 39,000+ university students) and 100+ on LeetCode
- Built multiple deployable ML systems across NLP, Computer Vision, and RAG
- Published Kaggle datasets with 60+ downloads
- Advanced Retrieval Augmented Generation (RAG)
- MLOps and scalable AI workflows
- Vector Databases
- LLM Applications
- Production-grade AI systems
- Email - bansalmayank1414@gmail.com
- LinkedIn: https://www.linkedin.com/in/mayank-bansal14/
- Portfolio: https://mayank149.github.io/Portfolio/