Fernando Scheinkman

I developed a code that downloads data from SAP. The client needed information from the past four months, which needed to be extracted from five different transactions. Some transactions were small enough to download the 4 months together, for others, it was necessary to download into 2 months, 1 month, or half a month segments.

I developed a system that downloaded this information in the most efficient way possible, choosing the combination of periods that would be more efficient. To achieve this, the bot first downloaded the five transactions in each possible way. Then it stores how much time it took in each combination. With this information, it proceeds to execute the code in the most efficient way. Every time the code executes, the robot stores how much time it took to download and then averages it against the information it had previously, and thus determine if it needs to change how it downloads the information.

I was a part of a major RPA project for an international retail company. The bot performed the cash count for 352 stores, allowing the client to reallocate the 15 employees that were involved with this project. The bot downloads excel files from a web portal, which are created by the stores' tellers. Then it inputs the cash register information contained in these excel file to another web portal. Once completed, it proceeds to execute the cash count in the web portal. If a discrepancy is discovered, it sends an email to the Sales Control team. Because of the volume of the data to process, the bot runs on five devices, four virtual machines and one notebook, using a queue. One of the major challenges in this process was logging each execution of the process. Because the client wanted access to this information, they wanted an Excel file. But because there were five devices trying to use this file, the process was both slow and prone to fail. So I decided to develop a small database, to which the client could access through an excel file, which solved the issue.

One thing I am proud of is that the bot not only inputs the data into the web portal, but it also controls that the data was correctly inputted, avoiding duplicates.

A UiPath automation that, based on a text file, searches through five different portals, commercial payors, and government payors and returns with the claim status. This process will generate an output text file to be uploaded to a healthcare ERM.