No better time to be an Application Architect than now

It is a fascinating and exciting experience to build an application from scratch. Many of us always wanted to be part of ‘AD’ (aka Application Development) project. In the early 2000s, most of the applications were being built from scratch across the Enterprises, also called Bespoke Applications or Custom Applications and all of us have had a great time in building several applications, adopting newer technologies thereby in this process. We always loved the experience and our Technology Career Path.

Gradually, enterprise needs which originally met through custom/bespoke applications, were being catered to by Commercial Off The Shelf products/Enterprise Resource Planning software. Evolution of Software as a Service (SaaS) has further reduced the number of custom/bespoke Enterprise Apps being developed. At one point of time, COTS/SaaS products catering to the implementation of Line of Business Systems, and functional systems, have made Custom/bespoke Application at Enterprises to a very minimal level, barring very few cases, where it was needed to develop an application from scratch.

Thanks to the confluence of cost effective offerings in the Cloud PaaS/Serverless/Cloud Native, and fueled by the growth of AI/ML/NLP across the enterprises, there has been a clear trend of resurgence in the number of custom applications being built over the last few years. The table shown in post's caption depicts the need of Application Architecture skills across a variety of Enterprise Apps.

Here are few examples that substantiate this trend. 

• Instead of implementing canned reporting in a typical commercial BI Tool, develop a cloud native Serverless solution for running workloads like Reports Filtering, Querying, Display, Excel, PDF Generations. Such solution costed $50 a month even with thousands of users. Of course, for ad-hoc and complex reporting this may not be the right way.
• A raw NLP algorithm bundled into a full-fledged enterprise bespoke application through AWS Serverless Architecture. Raw AI/ML/NLP algorithm generally warrants a custom application to wrap it around, for its wide adoption and Application Architecture skills are vital for its success.
• Instead of implementing integrations via Mulesoft, develop a cloud native solution very quickly using PaaS features of AWS like Batch/Fargate, S3, SQS (OR) a Serverless Stack. Such solution was proven to be at par or cost effective than a Mulesoft based solution. No way meant to replace Mulesoft, but for simple integrations or where Mulesoft has limitation(s).