Ambition
An international leader in the energy and petrochemical sector sought to optimize the maintenance and upgrade process for its serverless functions on AWS. The business had developed hundreds of serverless functions over the years using Python and found the process of upgrading these functions challenging. Each upgrade required manual adjustments that risked system downtime, diminished customer satisfaction, and invited revenue loss if not executed perfectly. Delayed upgrades posed additional risks including data loss, security vulnerabilities, and damage to the organization's reputation. Furthermore, the absence of comprehensive documentation for many functions also led to inefficient utilization of cloud resources and elevated operational costs.
Action
To address the client's maintenance and upgrade challenges, Wipro built a user-friendly GenAI-based application using Anthropic’s Claude 2.0 large language model (LLM) in AWS Bedrock . Leveraging React for the front-end and Python for backend scripts, this application automates impact analysis and upgrades for hundreds of AWS Lambda functions across various AWS subscriptions. The process encompasses securing permissions, managing source code, analyzing and upgrading Lambda functions, followed by testing, validation, and performance optimization through AWS services.
Ambition Realized
By automating upgrades analysis, the client reduced analysis time by approximately 75%, enabling the development team to focus more time on tending critical backlog items. The automations also reduced overall deployment time by nearly 50%, streamlining processes such as development and testing, and increased developer productivity by almost 25%.
Leveraging large language models (LLMs), the client executed its Lambda upgrades seamlessly and efficiently, well before the software’s end -of -lifecycle (EOL) support, ensuring uninterrupted system operations. Most notably, the client reaped substantial cost savings, amounting to $1 million per year, approximately 10% of its overall cloud expenditure.
