Information Technology student with a strong foundation in AI/ML and Deep Learning, experienced in model training, optimization, and deployment. Adept at building end-to-end ML workflows, analyzing large datasets, and improving model accuracy, efficiency, and inference speed across real-world applications.
- Improved segmentation IoU by 6–8% using RL-guided patch attention.
- Built training and evaluation pipelines.
- Tech: Python, PyTorch, Reinforcement Learning
- Built Random Forest + XGBoost ensemble improving precision/recall by 12%.
- Feature engineering & selection.
- Tech: Python, Scikit-learn, Pandas, XGBoost
An AI-powered complaint classification system built using a Retrieval-Augmented Generation (RAG) pipeline. The system enhances response accuracy by combining semantic retrieval with generative AI.
- 🔍 Semantic search using transformer-based embeddings
- 🧠 Retrieval-Augmented Generation for improved classification
- 🎤 Voice-based input for user queries
- 🌐 Interactive deployment using Streamlit
- Python
- Transformers (NLP)
- FAISS (Vector Search)
- Streamlit
- Improved response relevance using hybrid retrieval + generation approach
- Efficient handling of user complaints with contextual understanding
An automated OCR-based document processing system that extracts structured data from scanned marksheets and converts it into machine-readable format.
- 📄 Extracts structured student data from scanned documents
- 🧹 Image preprocessing for improved OCR accuracy
- 🔧 Handles noisy and low-quality images effectively
- Python
- OpenCV
- Tesseract OCR
- Applied preprocessing techniques like denoising, thresholding, and resizing
- Improved extraction accuracy under real-world noisy conditions
- Converts unstructured documents into structured digital data.
- Oracle Cloud Infrastructure 2024: Generative AI Professional
- AI/ML for Geodata Analysis – ISRO
- Python (Basic) – HackerRank
- Java (5 Star) – HackerRank
- Cisco Networking / Cybersecurity Essentials
Multiple Internships — Web Dev, ML, Data Analytics, Cybersecurity
- Automated data pipelines & reproducible ML notebooks.
- Cross-team collaboration on deployment and testing.
⭐ Thanks for visiting my profile!