Indian Institute of Information Technology, Allahabad
B.Tech in Information Technology (2023 - 2027)
Relevant Coursework: Data Structures, Machine Learning, Web Development, Database Systems
- Responsive Design: Built a modern, mobile-friendly personal portfolio website showcasing projects, skills, and experience.
- Interactive UI: Utilized HTML5, CSS3, and JavaScript for smooth animations and an intuitive user experience.
- Scalable Architecture: Integrated best practices for performance optimization and SEO, ensuring fast load times and discoverability.
Tech Stack: HTML5 | CSS3 | JavaScript | Bootstrap
- Innovative Pipeline: Engineered a scalable AI solution to semantically group and retrieve clinical trial data from 450k records (495MB of textual bio-medical data) by merging multi-source inputs and extracting structured relationships using Gemma2.
- Optimized Data Processing: Leveraged Stella 1.5B embeddings with FAISS (0.8 threshold, avg. cosine similarity 0.857) to perform advanced node deduplication, forming a dynamic Neo4j knowledge graph.
- Graph-Based Recommendations: Developed a state-of-the-art recommendation engine using Neo4j GDS (Jaccard similarity) to deliver top-10 trial recommendations, ensuring cost efficiency, clinical relevance, and explainability.
Tech Stack: Python | Gemma2 | Stella | FAISS | Neo4j GDS
- Data Extraction & Synthesis: Engineered an AI-driven platform that extracts and synthesizes multi-source data (PDFs, Wikipedia, Google) to analyze food ingredients, define safe consumption guidelines, and recommend healthier substitutes.
- Geolocation & Mapping: Integrated geolocation-based store recommendations with interactive mapping using Google Maps API and Folium for enhanced user experience.
- Enhanced Semantic Search: Leveraged FAISS and SentenceTransformers to boost semantic search accuracy and deliver precise insights.
Tech Stack: TensorFlow | Streamlit | GroqAPI | Google Maps API | FAISS | SentenceTransformers
- AI-Powered Analysis: Developed a full-stack AI web application that predicts skin diseases from user-uploaded images and provides follow-up analysis and treatment recommendations.
- Dynamic Diagnostic Flow: Implements follow-up questions powered by custom logic and machine learning to refine diagnoses and increase prediction accuracy.
- Comprehensive Treatment Insights: Displays treatment solutions in a well-organized format:
- Ayurvedic Solution
- Home Remedies
- Over-the-counter (OTC)
- Prescription Drugs Each with concise 1–2 line explanations.
- Robust Architecture: Built with a Next.js and TailwindCSS frontend paired with a Python (Flask/FastAPI) backend, ensuring secure and efficient client-server communication.
Tech Stack: Next.js | TailwindCSS | Python (Flask/FastAPI)
- Interactive Web Application: Developed a user-friendly web app that converts sketches to colorized images in real-time using generative adversarial networks (GANs).
- Custom GAN Model: Trained a bespoke GAN on an anime dataset to achieve high-quality colorization with reduced latency.
- Seamless User Experience: Delivered an interactive, responsive interface ensuring an engaging and intuitive user experience.
Tech Stack: GANs | Flask | HTML5 | CSS3 | JavaScript
- Predictive Analytics: Developed a supervised machine learning model to forecast employee attrition by analyzing demographics, job satisfaction, work-life balance, and performance metrics.
- Data Visualization: Utilized matplotlib and seaborn to visualize data insights and model predictions, providing actionable intelligence.
- Robust Model Development: Built and trained the model using TensorFlow to support effective retention strategies.
Tech Stack: TensorFlow | matplotlib | seaborn | Python
- Comprehensive Campus Platform: Created a web-based platform to streamline campus management, enhancing the student experience with secure authentication, event organization, and resource allocation.
- Real-Time Connectivity: Integrated real‑time notifications and a robust feedback system to foster effective communication.
- Dynamic Content Delivery: Leveraged the Gemini API for dynamic content and utilized Postman for rigorous API testing, ensuring scalability and reliability.
Tech Stack: Node.js | MongoDB | Gemini API | Postman
- Conversational Finance Management: Developed an intelligent WhatsApp chatbot to simplify financial management through natural language interactions.
- Advanced NLU: Built using the RASA framework for sophisticated natural language understanding and integrated with the Twilio WhatsApp API for seamless messaging.
- Efficient Data Handling: Employed Hugging Face Transformers for zero-shot classification and NER, with SQLite ensuring lightweight yet robust data storage.
Tech Stack: RASA | Twilio WhatsApp API | Hugging Face Transformers | SQLite
- Interactive Puzzle Experience: Developed a Java-based puzzle game featuring a sleek, intuitive UI with a 1-Click Solution feature for instant puzzle solving.
- Robust Desktop Application: Built using Java Swing to deliver a dependable and interactive desktop experience.
- Engaging Gameplay: Focused on providing a smooth and enjoyable user experience for puzzle enthusiasts.
Tech Stack: Java | Java Swing
- Core Member, GDG IIIT Allahabad (AI/ML Wing)
- Contributor, OpenCode IIIT Allahabad (Open-Source Event)
NodeJS | React | Advanced Machine Learning
🚀 Turning ideas into impact with code.