• Built a multilingual RAG-powered chatbot using Node.js and Gemini, enabling users to access accurate, context-aware insights on clean energy companies.
• Designed and implemented a RAG pipeline by scraping a directory with 20 cleantech companies with Puppeteer, generating vector embeddings via Hugging Face, and storing 10,000+ vectors in a DataStax Astra database, enabling the chatbot to retrieve relevant answers in under 2 seconds and deliver highly accurate domain-specific responses.
• Integrated Google Speech-to-Text and Text-to-Speech APIs to enable real-time conversation in 50+ languages, supporting both voice and text input with instant language detection.
• Developed frontend features including voice playback controls (pause, play, repeat) and a clean multilingual interface, improving usability and making the chatbot accessible to non-technical and visually impaired users.

• Designing a feature using OpenAI API and web scraping to aggregate and summarize health articles, aiming to deliver 10+ AI summaries weekly for improved content accessibility.
• Planning a new React Native tab and article view to display curated health content with AI-generated insights, targeting a 40% increase in in-app engagement.
• Developing a dynamic content feed using Node.js and Flexbox layouts, enabling real-time article updates and mobile responsiveness across devices.
• Preparing integration of external article linking via the Linking API to allow users to open full articles in-browser, enhancing content usability.

• Used C to build a command-line tool (called gcal) for managing reminders, allowing users to add, remove, and view reminders for specific days.
• Used file I/O to save and load reminders from a file, enabling data persistence across 100+ program runs and preventing loss of reminders even in the event of crashes or interruptions.
• Developed a shell script (install.sh) to automate the installation and setup of the program, simplifying installation for users and cutting setup time by approximately 75%.

• Developed a recursive algorithm to implement the computer’s AI decision-making system, enabling strategic Tic-Tac-Toe gameplay for the user.
• Used a hash table with separate chaining to store game states (Win, Loss, Tie), which reduced collisions and resulted in a game simulation up-to 5 times faster.

• Developed a real-time currency converter API using Flask, supporting 150+ currency conversions.
• Implemented web scraping with BeautifulSoup and requests to extract live currency data, ensuring conversion rates are updated every minute.
• Delivered results in a structured JSON format, enabling seamless integration with third-party financial applications.