Sep 2017 – Aug 2019 · 1 yr 11 mos · Bangalore

• Worked on the Platform as a Service (PaaS) team within Data Platform at Visa. I was part of the Platform Analytics vertical in my team.
• My role involved working on an ETL pipeline to extract necessary logs/metrics, process them and present them in a form to provide valuable insights to the team for optimizations on the Hadoop platform usage as well as better, informed capacity planning for future.
• Specific use cases are -
• ❋ Estimating cost incurred by a Business unit for using our platform to run their Hadoop applications.
• ❋ Visualize major consumers in terms of storage/memory/processing for our Hadoop Platform.
• ❋ Detect anomalies in terms of small or extremely large HDFS file sizes.
• ❋ Centralised Hardware Repository to view hardware specifications regarding all servers in our Platform.
• Technologies Used - Elasticsearch, Filebeat, Logstash, Kafka, Spark, Cassandra
• Continuous Integration, Continuous Delivery(CI/CD) - Successfully
• completed integration of CI/CD (automation of project build till deployment)
• tools with our existing project - Hadoop Platform Analytics.
• ❋ Campaigned for the best practices of the same to be used across the
• team through coherent documentation and knowledge sessions.
• Also, I mentored a project to automate the process of onboarding applications to Hadoop Platform. This process enhancement reduces the efforts on the part of the external application team to raise multiple requests and also brings down the overall process time to one-third.

Jun 2017 – Aug 2017 · 2 mos · Geneva, Switzerland

• Selected among 37 students for Openlab Summer Student Program, 2017 at CERN from 1560 applicants.
• Data Generation for NextGen WinCC Open Architecture SCADA system -
• ❋ Developed a data generator to simulate actual data that would be generated by experiments at CERN in coming years to replicate the scenario and perform relevant testing in advance.
• ❋ Data was mocked in terms of data type, frequency of generation and start/end time of generation. Optimizations were brought in by using necessary sort and accumulate algorithms to boost performance.
• ❋ As a result, data points were generated at the rate of 5 million entries for a simulated time duration of a second.

May 2015 – Jul 2015 · 2 mos · Greater Hyderabad Area

• SQS Queue Accumulator & Batch Indexer in Elasticsearch -
• ❋ Implemented accumulation and indexing of time series data retrieved using Amazon Web Services (AWS) Simple Queue Service (SQS) in Elastic Search. The scheduler triggered multiple threads for accumulator and subsequently indexing action into Elasticsearch. Also, indexer supported features like synchronous/asynchronous indexing and single/batch indexing.
• ❋ Libraries like Gson, Guice, Guava’s Maps, Joiner, Stopwatch, Java 8’s Optional, Completable Future and testing frameworks like Mockito were consistently used to complete this implementation.
• ❋ Learnings/Technologies Used - Elasticsearch, AWS SQS, AWS SNS, Java 8 framework, Unit testing, Multithreading/Concurrency frameworks