• Architected and deployed a private Kubernetes cluster on Google Cloud Platform (GCP), designed to support custom data science environments with secure, scalable access within a specific VPC.

• Implemented OAuth authentication with Google Authenticator using Terraform, providing secure access control and managing OAuth credentials within a cloud-native infrastructure.

• Developed comprehensive Terraform scripts to automate the creation and management of the Kubernetes cluster, Google Cloud Storage buckets, and Network Endpoint Groups (NEGs), facilitating seamless and repeatable deployments.

• Part of the Alexa team that is helping to integrate LLM into the architecture

• Researching and testing out multiple LLM upon certain metric to find the best performance of each model with the help of Amazon Bedrock and Sagemaker

• Created an entire Flask Web Service that would be running an LLM model when a request comes to Alexa, and hosted all under an EC2 Instance

• Initiated security measures for external services using AWS SecretsManager (token authentication)

• Implemented an in-memory vector database that store relevant metadata to compare user request to, and produce best results based on model response

Part of a development team in reading splunk log lines, applying ML algos in determining the exception/error state patterns/severity/occurrence and publishing the server log analysis to downstream dashboards/reports for engineering reviews.

• Supported the re-engineering initiative to generate an manageable design / flow-chart underlying an huge C++ code-base.

• With extensive development in the last two years, this code-base had grown exponentially with limited abilities to understand the problem that it was trying to resolve.

• Reviewed the code-base and its multiple dependencies in arriving to an design with which future enhancements will be more organized.