Oct 2020 – Aug 2021 · 10 mos

• Worked on solving the following business use cases of matching with help of a generic matching engine:
• 2A vs Purchase Register GSTIN level Matching
• 2A vs Purchase Register PAN level Matching
• 8A vs Purchase Register GSTIN level Matching
• GSTR1 vs Sales Register GSTIN level Matching
• 2B vs Purchase Register GSTIN level Matching
• GSTR1 vs EwayBill GSTIN level Matching
• Worked on providing different post-matching functionalities on matching data like viewing, filtering, sorting, pagination, single/multi-select/bulk actions, report generation (CSV and Excel), vendor communication
• Worked to provide support for Eway Bill ingestion in GSTR1 vs EwayBill matching
• Built Recon as a Service for the seamless integration of ERP customers
• Other products: Matching-lite, Max-ITC

May 2019 – Oct 2020 · 1 yr 5 mos

• Started with working on the core B2B/B2C products of ClearTax (E-way Bill and GST) as a part of Enterprise Team
• Developed some really useful features for Eway Bill product including "Automated escalation of emails before Eway bill expiry implementing distributed locking using Redis to trigger task only from a single EC2 instance"
• Planned, Designed, and Integrated (View Only, View+Download) Access Control feature (ACL) at Business and GSTIN Level for enterprises in Cleartax GST Product
• Implemented Custom Role Management Functionality (ACL) where a user can create custom roles for himself providing different permissions at the module level
• Reduced loading time for reconciliation (matching of 2A and Purchase docs) guide page by breaking down the DB query into smaller queries and adding right indexes on the table. Hence, decreasing the number of row scans for each query, parallelizing all calls, and aggregating output in the application itself. This helped in reducing the overall latency of API.
• Fixed several bugs in the GSTIN Verification module and added support for multithreading which helped in reducing the overall report generation time by making GSP calls parallelly
• Helped in reducing the failure rate of Advanced Recon to less than 0.01% by fixing issues through the events/alerts raised during exceptions
• Designed and implemented Mentat(generic matching engine) consumer application from scratch. Built using ExpressJs, the idea was to consume messages from SQS and run matching on LHS and RHS docs on the basis of certain rules provided. CI/CD pipeline setup using Jenkins, Docker, and K8s.
• Triaged on-call tickets, identified, reported, and fixed a few critical bugs that were present in old recon.
• Tools & Technologies Used : Java8,DropWizard,Hibernate,MySQL,Docker,K8s,Grafana,Metabase,Retool,Pivotal,Scalyr, Swagger,NewRelic,AWS(SWF,S3,RDS,SQS),gh-ost,Jenkins,Spinnaker, Redis,JUnit,ExpressJs,Typescript,Jest,JIRA,Octant,Terraform, Percona,JMeter

Jul 2017 – May 2019 · 1 yr 10 mos · Noida Area, India

• Was part of EVER team in System S/W Group at SRI Delhi (Visual Display Division)
• Worked on the maintenance and stabilization of existing modules/tools/daemons present around ensuring robustness and stability in the target(VD, TV, LFD)
• Built new features to enhance the capability of daemons running which helped in identifying critical performance-related issues like memory leaks, overconsumption of flash(disk) space, excessive CPU usage, etc. at a global level
• Also built a tool for better visualization and performance monitoring of all apps, daemons running in the target at the application layer. It helped to monitor processes in terms of vital performance parameters like CPU, Memory(RSS and Swap), Flash, Temperature, etc. and provides information up to thread-level by importing a history snapshot containing all parametric information bound with a timestamp extracted using daemon running in Smart TV. This solution helped as a substitute for the very famous DS-5 Streamline tool which is licensed software.
• Helped Samsung in doing a lot of cost optimization by building a solution that generates a status code to be used by service engineers to identify the root cause and probable solution to a problem being faced by the customer on their Smart TV.
• Responsible for the complete design, implementation, commercialization of backend daemon of TV Device Manager Application, and integrating it to frontend Tizen web application.
• Was also involved in the development of other Tizen Native and Web Applications for Smart TV
• Tools & Technologies Used: C/C++, JavaScript, Python, Perforce, DML/DUMA, GDB, JIRA, SQLite, Tizen App Development, Splunk, Chart.js, Plotly.js
• Achievements:
• Achieved Professional SWC level Competency at Samsung
• Project of the year award 2017 (EVER Team)
• 2nd best code reviewer of month across the whole org unit

Jun 2016 – Jul 2016 · 1 mo · Bengaluru Area, India

• My role was to work in a team to completely redesign and revamp the e-commerce web app of Tailorkraft (to sell customized shirts) from scratch using a free and open-source JavaScript web framework "MeteorJS".
• I was involved mostly in the back-end part. Worked on MeteorJs, Mailgun, MSG91, Instamojo, SemanticUI, MongoDB
• In the end, we made a large part of the code open source as Meteor Ecommerce on Github.
• Github: https://github.com/meteor-ecommerce/meteorecommerce

Mar 2016 – Mar 2016 · 0 mo

• Reported minor bug fixes and issues in jsgte (Open source tool written in javascript for doing computations in game theory)
• Contributions: https://github.com/gambitproject/jsgte/pulls/PARAG00991

Nov 2015 – Feb 2016 · 3 mos

• Fixed several bugs, raised feature requests, and reported few issues in the Processing IDE (An IDE built for the purpose of teaching the fundamentals of computer programming in a visual context)
• Contributions: https://github.com/processing/processing/pulls/PARAG00991