NO PLACEHOLDER BIDS. YOU MUST READ THE SCOPE OF WORK IN DETAIL, AND ENSURE THAT YOUR BID INCLUDES BOTH PHASES OF THE PROJECT --------- DO NOT BID WITHOUT THE PROJECT PASSWORD --------- PROJECT COMPLETION IS 7 DAYS. -------- We require the development of an application that uses machine learning for incoming invoices and receipts. Combining regression, classification, and clustering, the output is to establish a repository of purchases and create predictive analytics for future transactions. All invoices and receipts are received via email and the transactions need to be saved into the database. Scalability is integral to the design, as we expect to process over 100,000 transactions per day. To ensure you have read this scope, please include the word 'Kitty' in your bid, to be considered. Background We originally created this project in Laravel PHP, and there is an existing work in progress to be able to view the expected output. In the current version, a ‘template’ was created for each merchant. Incoming emails were parsed and the content was grouped into the following sections: Merchant Details: General information including Address, Store number, and any contact information. Transaction Details: Transaction number, time, date Product/ Service Details: Items that were purchased as part of the transaction. Payment Details: Method of Payment, Reference number, and Amounts The limitations of the current solution was that it required constant updates on Merchant templates, and was time intensive with regards to creating templates per merchant. We began working on the Machine Learning project earlier this year, and have a partial work in progress but the developer had to terminate the project due to illness. We can provide the files. Machine Learning Initiative Our decision to rebuild the project using Machine Learning, is based on the important goal of scalability and expediting our schedule to the beta version of our application. Using ML, our expectation is for incoming invoices / receipts to be quickly allocated to the correct tables in the database including categorization. We also expect the application to identify the products/services for the purpose of building a universal product/service database. Baseline Datasets We can provide several different items to serve as baseline data sets for training the AI. These are: Our current database from previously parsed invoices Sample emails from multiple merchants and point of sale applications Scope of Work: This PROJECT WILL HAVE TWO PHASES PHASE 1: Use OpenAI to process incoming invoices, and get to MVP state Detailed scope of work here: https://docs.google.com/document/d/1NL1iiFs6f7CCnySfP_gs9hHbZGwb5c5aeyLycOjjQ9I/edit?usp=sharing PHASE 2: Integrate saved data into our database. We have an existing platform where we can review the parsed data Save data to Db We will access our platform and check to see if your saved data matches the incoming email receipt. Project Password: To ensure you have read this scope, please include the word ‘MINDEN’ in your bid, to be considered