Prerak Mathur

Incorporated Cost Attribution service into around 11 major production clusters, providing a bridge between device cost and cloud service cost, which is pivotal for determining device licensing costs.

Orchestrated development of Chaos Engineering platform currently deployed on 3 large research clusters, providing real-time cloud failure simulation.

Led the design and integration of multiple failure types, including infrastructure and service failures, and enabled multiple validations to ensure comprehensive testing, enhancing the resilience and fault tolerance of Aruba cloud and making it more robust against potential disruptions.

Created "econvert" Python library to convert monetary and computer size values in Excel/arrays to float/string formats, added CPU conversion functionality, and developed multiple scripts for data manipulation and meaningful information extraction from Excel/CSV/JSON files.

Authored a Python script for PostgreSQL shard analysis, including customer count, device tracking, overload detection, migration optimization, and activity status.

Built a multi-layered AI based recommendation system to identify client device performance issues ahead of time and offer them accurate diagnoses and remedies.

Developed a single deep learning model that would eventually replace 66 different machine learning models for assessing the severity of a system's performance, eliminating substantial overhead in the machine learning pipeline.

Developed deep learning models identifying the root cause of performance issues of devices with an accuracy of 92%.

Developed a script automating the process of retrieving data from AWS S3 storage while also performing detailed EDA.

Worked collaboratively with other teams for the migration of the application to microservices architecture.

Led a team of 11 people to develop program routines to monitor environmental parameters and transmit the data to the host server every 3 minutes.

Developed interactive web pages and forms for over 20,000 users to input data. The data was used for developing a web mapping service named "Bhuvan”, which facilitated a 2D/3D exploration of India.