Portfolio: Python - AI Agents, Data Pipelines, API Integrations And More

Upwork

Usama Navid

View Usama N. profile on Upwork, the world’s work marketplace. Usama has completed 25 jobs on Upwork. Check out the complete profile and discover more professionals with the skills you need.

PDF Parsers

School Prospectus (PDF) Fee Info Parser

A FastAPI application that processes PDF files containing language school information using OpenAI's GPT-4 Vision API. The application extracts structured data about courses, accommodations, and pricing. School Information Parser View Code

Parsing using OCR from PDF based Amazon BoLs from Emails into Google Sheet

An automated system for processing invoices and Bills of Lading (BOL) using OCR, with integration to SellerCloud and Google Drive. Features Automated email attachment extraction OCR processing of PDF invoices and BOLs Integration with SellerCloud for order management Google Sheets integration for data tracking Google Drive integration for document storage Scheduled processing with configurable intervals View Code

WoCommarce Invoice Parsing Solution and Inventory Management in Google Sheets

A FastAPI web application for an Ecommarce that automates the process of invoice data extraction and Google Sheet updating. The web application serves as a central hub for collecting and managing invoice data, providing both automated and manual options for PDF uploads. Created using FastAPI, OAuth2, and a variety of other technologies, the application is robust, secure, and highly automated. It features scheduled email fetching to automate the collection of invoices and offers data validation checks to ensure the consistency and accuracy of the populated data. Features User Authentication: Secure login mechanism to authenticate users. PDF Upload: Interface to upload PDF files manually. Data Extraction: Extract specific fields like order code, shipping address, product details, etc., from the uploaded PDF. Google Sheets Integration: Populate the extracted data into a Google Sheet. Scheduled Email Fetching: Automated function to fetch PDF attachments from a user’s email at scheduled intervals. Data Validation: Validation checks to ensure data consistency and accuracy. View Code

LLMs, AI Agents, Chatbots

Custom Chatbot in Fast API with Asynchronous Requests

FastGPT a high performance, customizable GPT chat application built with FastAPI and OpenAI's API. It implements several features essential to providing an excellent user experience, in addition to performance, and features a simple frontend UI built with HTML, CSS, and JavaScript Features 📝 Full conversation history functionality 🤖 Chat with the GPT-3.5 & GPT-4 models 🧰 Change system message in realtime from UI 💬 Real-time response character streaming functionality 🧩 Markdown support in both user and assistant messages 🎨 Syntax highligting for code blocks with automatic language detection ⚙️ Automatic scrolling with new messages, cancellable by scrolling up View Code

API Integrations And Syncs

Spike API Integration

Created a minimalistic health data application that leverages FastAPI for the backend, Google Sheets for data storage, and Spike API for health data collection. The application features a Dockerized environment and is organized into a modular structure with a clear separation of concerns for API, models, and utilities. It provides end-to-end solutions for user authentication, data collection, and storage in a seamless manner. The app is hosted on Railway, making it easily accessible and scalable. Deliverables Dockerized FastAPI backend application. Integration with Google Sheets for data storage. Spike API integration for collecting health metrics. User authentication and provider selection through FastAPI routes. A responsive frontend built with Bootstrap. A modular code structure with clear separation into API, models, and utilities. Makefile for simplified project setup and maintenance. Deployment on Railway for easy access and scalability. View Code

Automated Data Extraction from Shopware API and Reporting Solution for Bob's Auto Belmont

Developed a comprehensive data automation solution for Bob's Auto Belmont, a leading automotive repair shop. The goal was to streamline the extraction and reporting of key shop metrics, including daily and weekly email reports, to provide valuable insights and improve business operations. Deliverables Wrote a script to fetch data from the ShopWare API and extract relevant metrics such as upcoming appointments, closed sales, payment information, and technician billable hours. Created a daily email report with information on upcoming appointments, closed sales, payments, account receivables, product categories, technician billable hours, and parts margin below 40%. Generated a weekly email report that built upon the daily report, including upcoming appointments, closed sales and gross profit trends, and technician billable hours. Designed the email reports using HTML formatting for a visually appealing and informative layout.

Telegram Subscriber Count Bot

A telegram bot that counts and updates subscriber count of the given telegram channel in a google sheet Commands /start - Shows the current status of the bot and also shows some helpful tips. /admin - To mark a group as admin chat, can be marked once /auth - To authorize a google service account /sheet - To set up a google sheet /worker - To connect a telegram account into the bot /refresh - To manually trigger the member count task View Code

HouseCall Pro and Zoho Sync

This integration system provides seamless synchronization between Zoho CRM and HouseCall Pro platforms. It handles bidirectional data flow for customer information, deals, estimates, and job statuses through a webhook-based architecture. The system maintains data integrity with OAuth2 authentication and comprehensive logging to Google Sheets. Features Webhook handling for both Zoho CRM and HouseCall Pro events Customer data synchronization Deal/Estimate management Automated lead source tracking Google Sheets integration for data logging OAuth2 authentication for Zoho CRM Comprehensive logging system View Code

DHL-API to Get Pricing, Create Shipment, Get Labels From Google Sheets

These are some python classes to integrate DHL-API to Google Sheets Backend using Flask. The classes include DHL express API Class DHL Global Mail API Class Google Sheets API Class Google Sheet API Class for Development Google Drive API Class Google Drive API Class for Development The scope of the project was to automatically get Rates from the DHL API and update the Google Sheet with the rates. After getting the rates, the user can select the rate and the shipment will be created in the DHL API. The shipment details will be updated in the Google Sheet. The label from the DHL API will be automatically uploaded to Google Drive. The link to the label will be updated in the Google Sheet. View Code

Grubhub Orders API Reporting In Sheets

This project automates the process of managing employee and supervisor-related Google Sheets. It creates new sheets for each employee and supervisor, populates them with relevant data, and shares them with the respective users. The project utilizes the Google Sheets API and Google Drive API for sheet management and sharing. Features Automatically creates individual sheets for employees and supervisors Populates sheets with relevant data from a master sheet Shares sheets with appropriate access levels Handles duplicate sheet names gracefully View Code

Scrappers

Real Estate Websites Scrapper in Netherlands

This repository contains a Python script for interacting with the assets.nl website. It automates the process of logging in, filling in a form to create a new building, and extracting the building's price. View Code

KSA Lawyers Data Scrapper

A data collection and analysis project focused on lawyers' information in the Kingdom of Saudi Arabia. View Code

API Development

API to Generate PDF and Docx Resume for Adam Smith International Internal HR Product

A FastAPI-based web application for generating and managing CVs. Features CV generation from templates Google Cloud Storage integration PDF conversion support RESTful API endpoints View Code

Fast API Based Blog

FastAPI Blog 🔥 - a modern, fast (high-performance) blogging platform built with FastAPI and SQLAlchemy. This application provides a robust backend system for blog management, offering features like user authentication, blog creation, updating, deletion, and real-time blog updates. Features User Authentication: Secure login system to access token-based authentication. Blog Management: Create, read, update, and delete (CRUD) functionalities for blog posts. Real-Time Updates: Leverage FastAPI's capability for real-time communication. Data Validation: Comprehensive validation for user inputs to maintain data integrity. Scalable Architecture: Designed to be easily scalable for future enhancements. View Code

Full Stack Development

AI Newsletter Generator

The AI Newsletter App addresses this need by delivering timely and personalized content directly to users' inboxes according to their individual preferences. Features A secure platform for users to sign up and log in. Users can specify interests, update them as needed, set newsletter frequency, select preferred delivery time, and choose a time zone. The system automatically fetches relevant content from a designated API, formats it, and sends it at the user's preferred time. The application ensures that emails are sent at the selected time based on the user’s local time zone.

CRM in Django

This is a simple Customer Relationship Management (CRM) application built with Django. Features User Registration: New users can create an account. User Authentication: Users can log in and log out. Customer Records: Authenticated users can view customer records. View Code

Misc

Streamlit to Exe Deskstop Converter

Convert any streamlit app to an executable windows app. View Code

Stock Back Testing using Polygon API and Google Trends API

A powerful stock analysis tool that combines market data from Polygon.io with Google Trends search data to identify potential trading opportunities. This tool analyzes pre-market activity, gap-up percentages, and search trend changes to help identify stocks with significant movement potential. Features Combined analysis using both market data and search trends Pre-market volume analysis Parallel processing for date ranges Gap-up percentage calculation Market capitalization filtering Google Trends hourly analysis (4-6 AM PST) Customizable filtering criteria Excel report generation with conditional formatting Colored console logging Concurrent processing with rate limiting Trading calendar aware (handles holidays and market closures) View Code

Smart Investment Advisor Custom GPT Powered Chatbot
This project is a smart investment advisor application built using Python, Streamlit, OpenAI, and yfinance. It uses OpenAI's GPT-3 model to answer user queries and provide investment advice based on the latest stock prices fetched using yfinance.
Features
  • Fetches the latest stock prices using yfinance.
  • Uses OpenAI's GPT-3 model to answer user queries.
  • Provides a chat interface for user interaction.
  • Stores chat history in session state.
View Code
Pepsico Potato Chips Quality Control Using Computer Vision
A deep learning based approach to inspecting the quality of the potato using impages gained through a Jetson Nano enabled classifier.
Fashion Cloth Text to Image Search Engine
Text to image search ai app using haystack and clip
Use Cases
  1. E-commerce Platforms: When users want to find products similar to a reference image or description.
  1. Media Libraries: To search for images or videos based on textual queries or vice versa.
  1. Research: For tasks like object identification in images based on textual descriptions.
Containerized Transcription API using Whisper Model and FastAPI
This project sets up a robust web application using FastAPI. It integrates the open-source Whisper AI model for Speech-to-Text (STT) transcription, enabling accurate and efficient audio-to-text conversion. The application is containerized using Docker, ensuring consistent and efficient runtime in any environment.
Features
  • FastAPI Backend: A high-performance, easy-to-learn backend.
  • Whisper AI Integration: Cutting-edge STT transcription.
  • Dockerized: Simplified deployment, scalability, and management.
Fast API Using YOLOv3 And YOLOv3 Tiny For Detecting Common Objects
Made with