Projects & Portfolio

A curated view of featured work and upcoming systems I plan to build.

Text-Organizer Desktop App (Linux)

Completed

A lightweight Linux desktop utility for transforming text through simple construct, strip, and split operations. Designed as a fast, offline-first productivity tool with native Linux packaging.

Key Features:

  • Construct text from comma-separated input
  • Strip leading and trailing whitespace per line
  • Split text into one-item-per-line output
  • Native Linux GUI with no internet dependency
  • Packaged and distributed as an AppImage

ℹ️ Note: AppImages may not run on Ubuntu 25.04 due to FUSE limitations. Works correctly on Ubuntu 20.04, 22.04, and most Linux distributions.

PythonLinuxAppImagePyInstallerlinuxdeploy
GitHub

AI Resume Analyzer

Completed

An AI-powered full-stack application that analyzes resumes against job descriptions and returns a match score, missing skills, and actionable improvement suggestions. Focuses on real-world AI system design using LLM APIs rather than academic model training.

Key Features:

  • Resume-to-job matching with structured AI output
  • LLM-powered analysis with enforced JSON responses
  • Backend request validation and orchestration with FastAPI
  • Clear separation between frontend, API, and AI logic
  • Production-ready architecture with independent deployments

⚠️ Note: The backend runs on a free-tier Render instance and may take 30–60 seconds to respond on first request due to cold starts.

Next.jsTypeScriptFastAPIPythonPydanticLLM APIsVercelRender
GitHubLive Demo

Customer Service AI Agent

Completed

A LangChain-based customer support agent that combines retrieval-augmented responses from a local knowledge base with optional API tooling for order and customer workflows. Includes both CLI and Streamlit interfaces for local and browser-based usage.

Key Features:

  • Tool-enabled agent with knowledge search, order status checks, and escalation flows
  • Retrieval-augmented answers powered by Chroma vector search over local support documents
  • Conversation memory preserving the last 10 exchanges for better continuity
  • Optional REST API integration for live order, ticket, and customer lookups
  • Dual interfaces with CLI mode and a Streamlit web chat deployment on Render

⚠️ Note: The live Streamlit deployment runs on Render and may take a short time to respond on the first request if the instance is idle.

PythonLangChainOpenAI APIChromaDBStreamlitRequestsRender
GitHubLive Demo
Page 1 of 2