Final Year B.Tech CSE Student | Data Science Enthusiast
- Passionate about Data Structures & Algorithms, Data Science, and Full-Stack Development
- Experienced in Python, Java, SQL, and Web Technologies
- Interested in Machine Learning, Topic Modeling (LDA), and Big Data Tools like Hadoop & Hive
- Currently exploring Advanced DSA and Scalable Data Processing
- Internship: Data Science Intern @ Celebal Technologies
Languages
Frameworks & Tools
A Natural Language Processing project that detects and classifies hate speech from text data.
Highlights:
- Used TF-IDF vectorization and Logistic Regression for classification
- Preprocessed datasets to handle imbalanced classes using SMOTE
- Achieved high F1-score through fine-tuned model hyperparameters
Unsupervised learning project applying Latent Dirichlet Allocation (LDA) to the 20 Newsgroups dataset for topic discovery.
Highlights:
- Implemented text preprocessing: tokenization, stopword removal, and lemmatization
- Applied Gensim’s LDA model to extract coherent topics
- Visualized results using pyLDAvis for better interpretability
A Retrieval-Augmented Generation (RAG) chatbot that answers intelligent queries based on real loan applicant data.
Highlights:
- Ingests and chunks structured loan datasets for semantic search
- Builds vector search index using TF-IDF + FAISS
- Retrieves relevant data chunks and uses Gemini 1.5 Pro for human-like answers
- Deployed via Streamlit for interactive user experience
Tech Stack: Python, Streamlit, FAISS, TF-IDF, Google GenerativeAI (Gemini), Pandas
An improved word-level autocorrector that suggests accurate spelling corrections using advanced NLP techniques.
Highlights:
- Integrated spaCy for lemmatization and NLTK corpora for real vocabulary coverage
- Ranked suggestions by Brown corpus frequency for realistic recommendations
- Edit-distance-based candidate generation with robust error handling
- Vastly improved over v1.0, which relied on a limited custom vocabulary
Tech Stack: Python, spaCy, NLTK, Brown Corpus, Edit-Distance Algorithm
An Android application designed to connect entrepreneurs with mentors and foster collaboration.
Highlights:
- User profiles for entrepreneurs and mentors
- Real-time chat and messaging system
- Article posting and discussion threads for knowledge sharing
- Clean and user-friendly Material Design UI
Tech Stack: Java (Android), XML, Firebase (Authentication & Realtime Database)
|
|
|
"Code. Debug. Learn. Repeat."
