About Me
👨💻 I have a deep passion for exploring the intersection of mobile development and artificial intelligence.
🎬 Film and storytelling are my great loves, with Sci-Fi and Biopics being my favorite genres. Currently, "Blade Runner 2049" tops my list of favorite movies.
☕️ As a coffee enthusiast, I take delight in using the WDT tool. My favorite drink is a rich, flavorful Café Cubano.
🔧 I enjoy hacking tools to tackle small problems and thrive on networking with like-minded individuals within the startup ecosystem.
Contact Information:
📧 Email: yashv.khaitan@gmail.com
📞 Phone: 925-660-9495
📍 Location: Dublin, CA
- Spearheaded development of "AffectEcho", an emotion translation model utilizing a Vector Quantized (VQ) codebook to capture and replicate five distinct levels of emotional intensity in speech to enhance the emotional accuracy of synthesized audio.
- Optimized neural network performance by fine-tuning the Adam optimizer, resulting in 15% faster training efficiency and 10% higher model accuracy.
- Collaborated on integrating the model with text-to-speech systems to enable effective cross-language emotion transfer while maintaining speaker identity.
- Developed a comprehensive month-long curriculum to teach 100 students JavaScript & HTML/CSS.
- Emceed “lightning talks” where students showcased projects they developed under the curriculum.
Python, Flask, Pandas, ChromaDB, Docker
A Retrieval-Augmented Generation (RAG) application using LLM techniques for personalized advice on academics and campus life at Purdue. Leverages past student experiences through forum discussions, indexing over 100,000 documents in a vector database to enhance accuracy and optimize retrieval processes.
Try it outSwift UIKit, Google Firebase, CocoaPods
A mobile app for organizations to post flyers of events @ Purdue. Leading a team of dedicated students to develop the app. Currently being used by over 200 students.
Try it outTypeScript, React.js, Next.js, Vercel
An AI-powered recommendation platform for Purdue students to make healthier food choices at dining courts. Developed on top of OpenAI’s chat completion API.
Try it outReact.js, Node.js, MongoDB, Google Firebase
A website marketplace built with the MERN stack, for Purdue students to buy and sell textbooks. Features advanced search filters, notifications for transaction updates, and a user-friendly interface.
Try it out- Programming Languages: Python, Java, JavaScript/TypeScript, Swift, SQL
- AI & ML: PyTorch, Pandas, LangChain, Sentence Transformers, Embeddings, Semantic Search Algorithms, Large Language Models (LLMs)
- Web Development: Flask, React.js, Node.js, Next.js, Express.js
- Mobile Development: Swift UIKit, CocoaPods
- Data Management: Vector Databases (ChromaDB), MongoDB, MySQL, Firebase
- Cloud Platforms: AWS (S3, EC2), GCP (Compute Engine, Cloud Run, Cloud Scheduler)
- DevOps & Deployment: Docker, Heroku, Vercel