• Building AI agent pipelines to scrape websites, process data via AI prompts, and deliver results to the frontend
through REST APIs, powering UGC generation for 50+ websites
• Integrating Supabase to manage and retrieve AI-generated content, streamlining data flow across image, user,
and catalog systems and reducing content delivery time by 35%
• Managing deployments using Vercel and Railway and implementing CI/CD pipelines to streamline releases
• Developing a responsive React.js/Next.js frontend with Tailwind CSS, implementing dynamic product pages
and API’s across pages

At Semantic Technology Services, I’m working on building tools to automate the collection and analysis of sports performance data. My work involves designing Python-based pipelines using Selenium and pdfplumber to extract and process structured game data from PDF reports, significantly reducing manual data entry. I’ve also applied statistical methods and linear algebra to help analyze individual player contributions within dynamic, in-game systems. Additionally, I’ve engineered clean, analysis-ready datasets with pandas, NumPy to support data-driven evaluation of team and player performance.

As an intern at Lucid Data Hub, I actively participated with experienced & senior team members to support Gen AI application development activities, collaborating with a team of skilled professionals to gain hands-on experience in the space of data analytics.
My responsibilities included:
- Writing Python code to solve data analysis profiling challenges.
- Learn new skills and tools.
- API documentation.
- Participate with team in design discussion and product brainstorming.

This project is a local web app that gives users the ability to generate clean, aesthetic HTML-only clones of any public webpage. The application uses Playwright to completely render the target website, then scrapes the full HTML and CSS from that website. The extracted HTML are parsed with BeautifulSoup, while a local LLM (CodeGen-350M) regenerates simplified production-ready HTML. The backend is implemented with FastAPI, and the frontend is implemented with either Next.js or React, giving users a seamless experience.
This tool showcases how AI can facilitate and simplify the website replication experience while mitigating accessibility and ensuring code quality.

This project is an accurate model to identify hazardous asteroids, achieving 95% accuracy and 98% recall with ANN. This was done by constructing three distinct models including Perceptron, Logistic Regression, and Artificial Neural Network (ANN) for classifying hazardous asteroids.

SnapScan is a project I am currently making in my engineer course to help visually impaired users identify medicines, labelled food items and food items without labels. By scanning a product, users can receive additional information, including nutritional facts, allergy warnings, and other essential details. The technologies used in this project are python, tensorflow/keras for training the food and medicine model, OpenCV for processing the image and tkinter for the front end.

The Ishihara Test Program is designed to help users test their color vision, more precisely their ability to distinguish between red and green colors. The program simulates the popular Ishihara color vision test, which consists of a series of images with colored dots that are arranged in such a way that individuals with normal color vision can perceive numbers or shapes, while individuals with color blindness may not be able to see them.

This project was part of my engineering course, where I and my team were tasked to create an efficient mechanism to transport luggage and software to help maintain, track and organize airport and flight data. The software part of the program efficiently analyzes large files of fleet data, retrieves useful information and displays it to the client in a visually appealing way. For further information visit: https://able-fridge-faa.notion.site/P1-1P13-International-Airport-16dd75a511d78072b850d7a0c19787ec

MemoAR provides users with a seamless narrative of any statute, allowing you to Structure-ify your knowledge with our AR prototype! When users are faced with a statue and are unsure of the story behind the respective figure they are able to scan a QR code located on the statue, bringing them into an augmented reality narration, providing them with a backstory, fun facts, and other key elements of the figure's life.

BYTEsense is an AI powered tool which personalizes itself to a users individual dietary needs. First, you tell the app what foods you ate, and rate your experience afterwards on a scale of 1-3 - The app then breaks down the food into its individual ingredients, remembers how your experience with them, and stores them to be referenced later. Then, after a sufficient amount of data has been collected, you can use the app to predict how NEW foods can affect you through our "How will I feel if I consume..." function! This project was submitted for Hack The North 2023 and was awarded the 2nd best replit hack. For more information here's the devpost link: https://devpost.com/software/bytesense