AI/Agentic AI Projects

AI & Agentic Apps

Here are some of the AI applications and projects I’ve worked on, from Streamlit prototypes to full-stack TypeScript and Next.js builds.

AI Apps

TripBites

A travel assistant built with Next.js and LangGraph. It uses a multi-step workflow to gather trip information and return a useful answer in one place.

Problem: Trip planning often means checking several sources and pulling the details together by hand.

What I built: Built the interface in Next.js and the workflow in LangGraph, using a ReAct-style approach to break requests into smaller tasks.

Engineering note: The project focuses on workflow orchestration rather than treating the app as a single chat prompt.

TypeScriptNext.jsLangGraphReAct
View project →

VoiceBuddy

A document-to-audio app that summarizes uploaded documents and turns the summary—or the full document—into narration using OpenAI TTS.

Problem: Long documents are not always convenient to read, especially when someone needs to review them away from a screen.

What I built: Built a Streamlit interface that accepts documents, creates a summary when requested, and generates narrated audio.

Engineering note: The app supports both summarized and full-document narration instead of forcing one workflow on every document.

PythonStreamlitOpenAI TTS
View project →

CarInsure Bot

A question-and-answer app for car-insurance documents. It retrieves relevant policy information before answering questions in plain language.

Problem: Insurance documents can be difficult to search when a policyholder needs a quick answer.

What I built: Built a Streamlit interface with LlamaIndex to retrieve relevant content before generating an answer.

Engineering note: The retrieval step keeps the response tied to the available policy material instead of relying only on the model’s general knowledge.

PythonStreamlitLlamaIndex
View project →