<PROJECTS_AND_PORTFOLIO>

TECHNICAL_PROJECTS

PROJECT_FILES

Sharepoint to RAG pipeline

[2026]

Built a scheduled ingestion pipeline on AWS Lambda (weekly sync) using the Microsoft Graph SDK to pull files from SharePoint into S3, connected to an AWS Bedrock Knowledge Base as the vector store powering an internal AI chat tool.

STACK:
 Python // Terraform // AWS Lambda
AWS S3 Buckets // S3 Vectors // AWS Bedrock

Stock MCP

[2026]

A Model Context Protocol (MCP) server exposing real-time market data and technical indicators to an LLM over a remote HTTP connection.

STACK:
 Python // FastMCP // Cloudflare

Zoom Transcript Summary Pipeline

[2025-2026]

Internal AI tool integrating with the Zoom API via webhooks and REST endpoints to generate custom summaries of Zoom calls. Users could either trigger summaries manually with custom prompts or configure an automated pipeline for recurring summarization. Built primarily for sales teams to surface insights from their calls, the tool included a UI for managing, reviewing, and regenerating summaries, along with a function-calling tool that let an LLM pull call summaries on demand. I also implemented a RAG (Retrieval-Augmented Generation) system that let the marketing team query across the generated summaries.

STACK:
 Python // Streamlit // PostgreSQL // Flask // REST APIs
AWS S3 Buckets // S3 Vectors // AWS Bedrock

Air Quality Monitor

[2026]

Built a real-time air quality monitoring dashboard using a Raspberry Pi Zero 2W and a Sensirion SEN66 sensor. Dashboard displays live air quality metrics such as Co2, VOC and particulate matter and also provides historical data trends.

STACK:
 Python // Flask // Javascript // HTML // CSS

Telegram Stock Alerts

[2026]

A Telegram bot running headlessly on a raspberry pi that delivers real-time alerts on a watchlist of stocks, covering price movements and technical indicators like RSI and MACD. Separate from the alerts, the bot sends periodic AI-generated reports on the user's watchlist, including position-sizing guidance based on user-defined parameters.

STACK:
 Python // C++ // sqlite
  • > Repository includes a C++ data analysis engine used for calculating sharpe ratios.

StockChat

[2025]

Financial Tech application that displays real time Stock and Options data to users. Users can search a specific stock ticker, and generate an AI overview based on the stocks historical price, financial data, and latest news. Users can also select a specific option contract (date and strike price) and receive an AI overview of the contract.

STACK:
 Python // Flask APIs // React Native Javascript
Openrouter // sqlite // Azure
  • > Currently closed source.