Interactive Brokers Group - SD Intern June 2019 - Dec 2019 • Developed Batch monitoring scripts to support CAT application such as to identify audit processes and monitor their existence in Kafka cluster, report Audit collection processes as targets to Prometheus to monitor, monitor alive status of Kafka producers running on different hosts across the world using Python3. • Created a REST service to fetch the closing price of a company’s stock using Java Spring MVC, JUnit and Mockito. • Monitoring Apache Kafka which is used for Streaming Engine Using Prometheus and built dashboards using Grafana. • Developed a custom Prometheus Exporter using Prometheus Java client library in order to expose custom metrics to Prometheus. • Developed a mechanism to report FATAL errors occurred in Streaming Engine to rsyslog using log4j.
• Bitcoin transaction history analysis Sep 2018 – Dec 2018 – Crawled bitcoin transaction data, in JSON format, from a block in Blockchain using blockchain api and Beautiful Soup packages in Python3. – Analyzed the number of bitcoins sent, received and available with the user and visualized using plotly package. – Visualized particular transaction of a user by drawing a directed graph using networkx package in python.
• Libratus- Research Project Aug 2018 – Nov 2018 – Report on the Poker AI bot Libratus is made explaining game theory and difference between perfect and imperfect games along with the algorithm and techniques the AI bot used.
• Linux Scheduler Profiling Aug 2018 – Nov 2018 – Designed various benchmarks to measure overheads, latencies, runtime and behavior of different schedulers in a Linux machine and profiled them using tools like Kernel Shark, trace-cmd, perf, stress and stress-ng.
• Analysis and classifications of proteins critical to Mouse model learning of Down Syndrome Jan 2016 – May 2016 – In this work we have discussed various Dimension Reduction techniques for pre-processing the network intrusion datasets. – Two new algorithms namely, Ensemble Ranker and Wilk’s Lambda Score have been proposed and proven to be more effective means of Dimension reductionality. – This work was conducted on a dataset namely data_cortex, 2015.