- Leading the CRM software development- Handling 5 projects related to CRM for different companies which included ZohoCRM, Salesforce and EspoCRM.- Dockerizing the CRM applications and creating deployment strategies using Kubernetes for CI/CD- Building unit tests for the CRM projects- Configuring and optimising MySQL and PHP-FPM perfomance to match the resources available.- Integrate UniCourt with CRM through API's.- Standarized client integrations to CRM using Swagger API, OAS and Swagger Codegen.- Building code across several projects of different teams (website, web-application, and CRM) to make CRM more usable for a sales and support agent.- Built a Zap application prototype for clients to connect with UniCourt- Build Zap's in Zapier to integrate several third party applications like Zoho, GoTo Meeting, Acuity Scheduling, Google Calendar and AutoPilot.- Integrate Zoho with Nextiva using Go Integrator to allow click-to-call feature for the sales team.- Zoho was inadequate because it was not customisable enough. So, had to come up with a different open source CRM that would satisfy the needs of UniCourt. Integrating UniCourt with EspoCRM through API's.- Understand the architecture of EspoCRM which was built in Backbone.js.- Used Salt Stack to automate setting up AWS EC2 instance for EspoCRM.- Migrate from Zoho to EspoCRM with all the customisations done prior to EspoCRM and new customisations for UniCourt to satisfy all the needs of sales team.- Allowing click-to-call using Twilio and SIP softphone through VoIP in EspoCRM.- Integration of creating support tickets.- Writing automated workflows and Business Process Manager Workflows to automatically add leads to sales teams queue and remove from the queue when they are done with it.- Creating email templates to respond to support requests, to notify new lead creation/support ticket assignment.- Integrating EspoCRM with AutoPilot through API's to provide the clients an email journey with UniCourt.

Roles performed by Karthik in data mining team- Building extractors that would fetch and aggregate data from different sources, normalise the aggregated data and store them into database using python programming language.- Build/Upgrade API's using python's Django framework that would allow users to perform on-demand fetching of normalised data either from the original source or from the database.- Optimise and improve performance of extractors so that it would run quickly and consume less memory.- Build regression framework for the extractor to validate and verify correctness of the mined data.- Update salt stack to maintain the versions of the softwares used and automate setting up AWS EC2 VPC's.- Build and architect testing normalisation frameworks to be cost effective and time efficient by creating dry runs.- Scrapping websites that were built using javascript client-site redirection using Ghost.py with some magic to make the run-time was extremely quick. Note: python's urllib2/requests doesn't support scrapping sites that run through javascript. Ghost.py is headless browser and HTML mining through this is slow.- Architecting extraction frameworks that were cost sensitive.- Re-architecting DynamoDB table structure to make the document downloads performed by user unique thus avoiding re-download of the same document and migrating from the existing architecture to the new architecture. Document downloads are cost sensitive.- Mentoring new joiners, introducing them to data-mining architecture through KT's, code reviews, learning new ways of programming from them were part of the work.Roles of Karthik in CRM Team- Understanding and adapting to new frameworks and projects which were written in PHP Yii framework.- Understand the existing CRM framework which was ZohoCRM- Rebuild/Rearchitect Zoho CRM to satisfy the needs of the company without effecting the CRM workflows. Building workflows, adding new fields to CRM and integrating with UniCourt.

Proposal:The idea was to establish a database of functions, provide the functions to machine learning system and for the system to figure out / generate with various degrees of confidence a program or a set of programs that will hold true for the given set of inputs and outputs. Karthik was a member of a 5 people team and was a Back End developer. Karthik was responsible for the system to figure out / generate with various degrees of confidence a program or a set of programs that will hold true for the given set of inputs and outputs. Technology stacks used by Karthik:1. Golang / Go Programming2. MySQL3. JSONMentor: Ganaraj P R - linkedIn profile: https://www.linkedin.com/in/ganarajpr