This project section will cover details around multiple projects that I either directly developed independently or collectively as a part of the development and delivery team at Academy Mortgage or for my own purposes related to my personal business activities.
Academy Message Bus API
My development and implementation of a messaging service handler for Academy Mortgage was based on the needs to improve performance and get database specific data back to a client application where the company client applications would not be directly connected to our internal network. The Encompass application is utilized for tracking the loan origination process and is a third-party application developed by Ellie Mae.
Through customized development we added a plugin that allowed us to make data request calls securely through the Azure Messaging services and then return the results back to the client application for further processing. This customization reduced our typical client wait time from 60 to well over 120 seconds down to under 15 seconds for most requests.
Amy Email API
As a part of improving process and decreasing the time it takes from loan application to close. We now leverage various automation tools through a vendor to decrease human error and increase the speed to delivery while providing greater value to our borrowers. One of difficulties that our vendor experienced was attempting to use Outlook as a part of their UI based Robotic Process Automation (RPA). There are a number of limitations to RPA in general including the brittle nature of point and click within a UI and the time it takes to perform those operations relative to the application (Encompass) it needs to interact with. This ultimately is no different than what many QA engineers use when leveraging technologies such as Selenium to perform web-based UI Automation testing.
When our vendor attempted to expand the number of RPA client servers that operated Outlook on them they hit a limitation on the number of concurrent connected instances of Outlook using the same credentials to our O365 instance that was managing email communication to the RPA clients. To remove this limitation, I developed an API driven service to manage in and outbound communication. Not only for Email but for other methods of communication including SMS text, Encompass Event Messages, and Teams or Slack communication to or from the bot servers.
In my implementation I created an Azure Function App to act as the primary inbound message handling service or series of Webhooks for email and other requests coming inbound into the bots, including one Logic App that monitored the inbound email through O365. These inbound messages are then sent to an Azure Message Bus queue for later processing. On the other side I created a reader Azure Logic App that handled the pulling of messages out of the queue and other micro services for sending out email and other communication to clients and users. This also included sending out email messages with file attachments using Azure Blob Storage as the temporary file storage for those attachments.
Encompass (Dev Connect) Sync Service
This was primarily a proof of concept project to help develop replacement processes and technology of an aged data sync process using the older ESDK or Encompass SDK libraries and technology. Encompass by Ellie Mae has several services including Data Connect and Dev Connect to move data in and out of their Loan Origination Service. Dev Connect is typically used for the management of the data within Encompass performing various actions based on their Web API (Post, Put, Get, Delete). Whereas Data Connect provides a means whereby data can quickly and easily flow back to the customer in a secure and near real time manner providing the customer a SQL data store of their loan data to perform analytics and reporting.
The problem that this solution is to address is the replacement of the inbound data stream that we established using the Encompass SDK to pull data from Encompass into our own data warehouse. When this was developed, Ellie Mae had yet to provide a resource such as Data Connect, and we had to develop our own solution in house. In researching Data Connect, we really liked the solution but felt that the price point being asked was much higher than the value provided, primarily because we have multiple lanes in addition to our production environment raising the overall monthly costs beyond what we are ready to pay for as a company. Additionally, our aged Encompass SDK solution’s stability has always been in question and has required a great deal of maintenance and manual care to keep running on an annual basis.
To resolve this problem I created a proof of concept that leveraged the built in event based Webhook broadcast logic built into Encompass, and received those messages through a Function App micro service that then stored those messages in a Service Bus queue. Following a similar design to the Amy Email API solution I created a series of Webhook Listeners that Encompass could post messages to. These messages when received are sent to an Azure Service Bus for later processing when the Logic App puller is called. I then setup a puller service that makes a call to the Logic App event Web API event listener that then pulls a message from the Service Bus Queue and returns it to the service puller application. From there I then made a call into Encompass to retrieve the Loan Object by establishing a connection to Encompass and calling the Get Loan API call through Dev Connect.
This payload of JSON data is then stored in a simple database that I created to store both the original event request and the full loan data payload object. Conceptually from this point the Enterprise Data Team could read the data from the database and extract all relevant information from the JSON payload and store into our data warehouse for analysis processing and reporting for the executive teams.
- .NET Core
- Azure Function Applications
- Azure Logic Applications
- Azure Message Bus
- RESTful API
- Active Directory
- Java Web Token (JWT)