Business need
A Fortune 100 airline leveraged a legacy Teradata data warehouse to manage enterprise data, resulting in manual-intensive processes and spiralling operational costs. Troubleshooting was time-consuming and high failure rate significantly impacted operational SLAs. Additionally, legacy infrastructure made it challenging to build, deploy, and scale ML and GenAI use cases. The airline wanted to move to a scalable, reliable cloud architecture to enhance operational efficiencies, lower costs, and power AI-driven innovation.
Key business requirements:
- Modernize Teradata workflows to AWS, optimizing large-scale pipelines and infrastructure
- Ensure high data quality for downstream use cases with 24×7, end-to-end observability
- Simplify data operations, reduce manual effort, and lower operational costs
- Improve passenger experience with better platform reliability and deeper data-driven personalization
- Leverage ML & AI workflows to power chatbots, streamline baggage tracking, and enhance passport OCR, coupon-based offers, etc
Delivered a unified, agentic DataOps-driven platform, enabling enterprise‑scale ML and GenAI innovation for 3500+ users
Solution
Impetus established a reliable, scalable data foundation by modernizing the airline’s legacy data warehouse to a cloud-native data lake architecture leveraging AWS Glue, Amazon MWAA, Amazon Redshift, and Palantir Foundry.
Building on this foundation, the team implemented a unified AI/ML platform, which was leveraged by 3500+ users to host and manage 150+ business-critical applications. The focus was on controlling costs,
From CI/CD automation to proactive monitoring, the engagement enabled seamless automation, cost optimization, and operational excellence.
Solution highlights
- Seamless cloud orchestration: Migrated 100+ jobs to Harness, accelerated infrastructure deployment on AWS via IaC templates, while optimizing CI/CD pipelines.
- Automation and agentic AI: Optimized queries using SQL Ninja, accelerating troubleshooting with a specialized RCA bot, and enabled AI‑based review of MWAA DAGs and logs to prevent failures.
- End-to-end observability: Implemented a best-in-class observability platform for 24×7 monitoring and prevention of data quality issues.
- Best-in-class support & reliability: Delivered round‑the‑clock support for ML operations with proactive monitoring, incident and DR readiness.
- Governed MLOps: Established robust governance across IAM, AWS services, and GitHub, handling troubleshooting across applications and pipelines.
- Enterprise-grade security: Addressed 5000+ vulnerabilities across platforms and applications, and managed security compliance via Wiz and Veracode.
- Continuous optimization: Deployed proven optimization techniques to consistently control and reduce cloud spend.
Accelerated releases, with <10% change failure rate and 30% faster debugging, while reducing cloud spend by 25%
Impact
Impetus helped the airline accelerate their cloud modernization journey through agentic AI-driven DataOps, enabling deployment of ML and GenAI workflows at scale. This in turn empowered enterprise-wide users to innovate and unlock unmatched efficiencies across multiple use cases.
Business benefits
- Lower platform costs: 50% cost savings annually with data pipeline and platform optimization
- Faster issue resolution: 30% fast debugging for MWAA-based data pipelines
- Safer deployments: <10% change failure rate on Harness for AWS
- Higher AI adoption: 3,500+ AI-driven user requests served across the enterprise in 2025
- Improved risk management: 5,000+ crucial security vulnerabilities addressed
- Lower cloud spend: 25% AWS cloud cost savings achieved through continuous optimization
