Siddharth Thakkar

Built a full-stack AI-powered platform that generates and curates unbiased Wikipedia-style articles using GPT-4 and real-time web parsing, with automated fact-checking and editorial oversight.

• Engineered an AI-driven article generator using GPT-4 to produce topic-specific, fact-based content.
• Integrated real-time web scraping pipelines to enrich articles with current and reliable data sources.
• Developed a Python microservices backend for fact-checking, bias detection, and source verification.
• Implemented dual authentication (Google OAuth and credentials) with role-based access and rate limiting.
• Designed a community editing system with AI validation to preserve article integrity.
• Built the full-stack app with Next.js, Vercel AI SDK, Prisma ORM, and PostgreSQL, ensuring seamless integration with Python services.

Performed data-driven performance analysis of Real Madrid matches using Python to extract and visualize player and team-level insights.

• Scraped match and player data from public football stats websites and stored it in structured formats.
• Used Pandas and NumPy to clean, process, and aggregate match statistics across multiple seasons.
• Built custom visualizations to analyze key metrics like goals, assists, shot accuracy, and player efficiency.
• Automated generation of season summaries and comparisons across players and match types.

Developed a custom Model Context Protocol (MCP) server enabling Claude, an LLM, to execute real-time stock trades through Zerodha, an Indian brokerage, overcoming significant documentation scarcity.

• Engineered a custom MCP server as a bridge between Claude and Zerodha's trading API.
• Integrated deeply with Zerodha's developer API for authentication, order placement, and portfolio management.
• Defined a suite of 'tools' within the MCP server for trading actions, meticulously described for Claude's understanding.
• Implemented natural language processing to translate Claude's commands into precise API calls.
• Incorporated real-time confirmation mechanisms to ensure user control over trade execution.
• Navigated and solved complex integration challenges due to extremely limited MCP documentation for AI agents.

Developed a full-stack uptime monitoring application with alert functionality, secure user authentication, and efficient data handling across multiple apps using Turborepo.

• Implemented secure user authentication with Clerk for protected routes and data privacy
• Built modular frontend, backend, hub, and validator apps using Turborepo for efficient development
• Used Prisma ORM with PostgreSQL for robust uptime tick and validator data storage
• Designed a responsive monitoring dashboard using Next.js and Tailwind CSS

Built a real-time collaboration platform enabling over 50 users to work on shared projects and communicate instantly, with secure backend and responsive design.

• Developed real-time chat functionality using Socket.io for seamless user communication
• Engineered backend with Express, Prisma, and bcrypt-based authentication for data security
• Designed a responsive, cross-platform UI using React and Tailwind CSS
• Optimized backend logic and frontend responsiveness for scalability and speed

Built a full-stack learning path generation MVP integrating GPT-4o via Palantir Foundry’s AIP Logic, with secure authentication and a polished UI.

• Created an AIP Logic function in Palantir Foundry to interact with GPT-4o for generating structured learning paths based on user goals
• Built a React frontend with Vite, styled using Tailwind CSS, and enabled navigation via React Router
• Implemented authentication and input validation using JWT and Zod respectively
• Developed a backend using Hono running on Cloudflare Workers, with Prisma ORM for database interaction
• Configured CORS and default route redirection to ensure smooth frontend-backend communication

Developed automated stored procedures to streamline e-commerce operations, including product addition, inventory management, customer account creation, order processing, and product shipment tracking.

• Created reusable stored procedures for key business processes
• Designed efficient SQL queries to extract business insights
• Optimized business operations through transaction-oriented stored procedures
• Enabled data-driven decision-making with real-time insights

Developed a mobile platform for managing tuition classes.

• Integrated Firebase Authentication
• Used RecyclerView for efficient display of student data
• Implemented MVVM architecture
• Added features for assignment management and student progress tracking

Developed a Java-based warehouse management system.

• Designed Product and Tote classes for efficient management
• Developed Warehouse class to handle operations
• Implemented order fulfillment and tote merging functionalities
• Created a comprehensive test program

Developed a semi-autonomous quadcopter for human detection.

• Implemented ESP32 camera module for real-time video capture
• Designed a 3D-printed chassis using SolidWorks
• Built a custom flight controller
• Programmed flight dynamics using Arduino IDE

Developed a system to assist people with hearing impairment by recognizing emotions from audio inputs.

• Used Python, SVM, RNN, and CNN algorithms for audio analysis
• Employed Vox celebrity dataset for model training
• Built the system to identify emotional states and provide real-time feedback