• Created a Kubernetes-based job manager using the Sidecar pattern to run and monitor machine learning models to predict Enzyme Commission (EC) numbers along with a web interface and FastAPI application that has served 6000+ researchers to date. (https://clean.frontend.mmli1.ncsa.illinois.edu/)

• Built an interactive visualization/educational tool using AngularJS for chemists and students to construct molecules from building blocks as part of a NSF funded project to accelerate automated chemical synthesis and advance the discovery and manufacture of novel materials.(https://dmm.frontend.mmli1.ncsa.illinois.edu/)

• Took ownership of design and development of usage metrics and the reporting dashboard for a notification platform. Created ETL pipelines using Azure Synapse analytics and a PowerBi report showcasing the platform’s impact, which helped increase visibility and onboard more consumers.

• Developed a web application with AngularJS and .Net Core supported by SQL stored procedures to provide user management functionality to admin users, thus reducing engineering effort and allowing quick access provision.

• Created ARM templates for Service Fabric clusters and Helm charts for Kubernetes clusters. Configured and migrated various Azure resources like Azure Databricks, app services, service bus instances.

• Remodeled Linux and Windows-based applications into micro-services, deployed them in containers on a Kubernetes cluster, helped reduce Azure spend by $34K (28%) and considerably simplified the deployment process.

• Developed a web-based application in React to monitor and execute emergency actions on Autonomous Underwater Vehicles(AUVs) using an agent-based framework called Fjage.

• Designed and created a dashboard to simulate the path of AUVs and developed various other indicators to display processed data from sensors and actuators.

Designed Directed Acyclic Graphs using Apache Airflow to create data pipelines with Hive and Spark operators and contributed to building a cost attribution model based on CPU usage metrics on a heterogeneous cluster - allowing customers to monitor and reduce cost of their big-data queries.