Ping Ai

Ping is an AI-powered productivity assistant designed to help students and professionals stay organized without manual planning. It ingests unstructured inputs—such as text messages, academic deadlines, and LMS updates—and automatically converts them into structured tasks, calendar events, and reminders.

I built Ping end-to-end as a full-stack product, designing both the user experience and the agentic backend. The system uses LLMs to interpret intent, extract actionable information, and trigger workflows that sync with tools like Google Calendar and Notion. I implemented the automation layer using n8n, integrated Twilio for SMS-based interactions, and used Supabase for authentication and data storage. Ping reflects my interest in agentic systems, product thinking, and building AI tools that meaningfully reduce everyday friction.

Ping Ai - AI-powered productivity assistant

Mockly

An AI Live Voice Agent Interviewer — A low-latency web application enabling real-time voice conversations with an AI agent designed for high-fidelity interview simulations.

Technical Implementation: Engineered a bi-directional audio streaming pipeline using WebSockets and the Google Gemini Live API. Implemented custom Web Audio API nodes to convert browser microphone streams to 16kHz PCM data and decode incoming raw audio chunks for gapless playback. Built on React 19, TypeScript, and Vite for performance, with Tailwind CSS for a responsive UI. Integrated n8n webhook workflows for serverless form handling and data collection.

Mockly - AI Live Voice Agent Interviewer

Keeper.ai

Keeper.ai is a tool used to help travel nurses onboard to the perfect jobs for them. Our application also features a budgeting algorithm to help nurses plan out their budgets and tax filings accordingly. I created this by web scraping an online Job Board using BeautifulSoup to find appropriate listings for travel nurse jobs. I then Created, stored, and trained sample data on an ML model to predict contract costs and expenses for budgeting and tax optimization purposes, and finally, utilized Text-to-Speech API to create an AI dictation assistant that transcribes patient visits in real time and outputs relevant information as a PDF

EquityPlus

EquityPlus is an application that I created to help financial professionals create an AI-generated equity research report with the input of a company. I utilized web scraping tools via a Bing Search API to scrape relevant articles based on user input. I then leveraged Langchain and Pinceone via RAG to create an LLM designed to create an equity research report based on an inputted company. Finally, I implemented a Discounted Cash Flow valuation function for each given stock with financial python libraries and a regression model

ResearchPro

Research Pro is an application that I made utilizing a web search API and Natural Language Processing to help students find the right professors to reach out to for potential internships. I found that many students found trouble finding the right professors to reach out to, so I decided to build a search engine that would solve that problem for all high school students.

SrivastavaTrains

SrivastavaTrains is a marketplace of AI-based running coaches that I created utilizing Retreival Augmentated Generation (RAG) to create tailored AI-based running coaches for specific events such as the 5k, 10, or marathon. I fed these models data from the training logs of some of the best coaches in the world and consolidated them in a relevant pinecone vector database.