Historically, organizations have grappled with manual data gathering processes, hindering the translation of insights into action. In a strategic move towards enhanced scalability and efficiency, a leading Multinational Financial Services Company partnered with Infosys to leverage Google Cloud Vertex AI for ML model deployment capabilities.
The partnership aimed to transition data science models to the cloud, leveraging the advanced AI and ML capabilities of Google Cloud. By replacing manual processes with automated ML Ops services, the company sought to drive innovation and accelerate time-to-market for data analytics solutions.
Key Challenges
- Manual Data Processes: Existing manual processes required significant effort and time, hindering agility and innovation.
- Deployment Time: ML model deployment traditionally took 12-15 months, delaying time-to-insight.
Ready to experience?
TALK TO EXPERTS
The Solution
Revolutionizing ML Ops with Google Cloud Vertex AI
Infosys orchestrated the deployment of an advanced ML Ops platform utilizing Google Cloud Vertex AI, enhanced by the robust Infosys Cobalt ecosystem. This strategic integration facilitated a frictionless transition of existing models from on-premises and various cloud environments directly into the cloud sphere. The cornerstone principles for the platform were an event-driven architecture for dynamic load and transformation processes, alongside an intuitive low code/no code interface for streamlined AI development.
Comprehensive security measures were embedded across the pipeline, coupled with streamlined continuous integration and continuous deployment (CI/CD) pipelines, fortifying the deployment process. This transformative solution granted data scientists a more agile framework for ML model deployment, significantly boosting operational efficiency and fostering a culture of continuous innovation.
Accelerating Insights with Data-driven Decisions
- ML Ops Platform Setup with Google Cloud Vertex AI: Seamless migration and deployment of ML models, driving agility and efficiency.
- Data Insights and Recommendations: Customized dashboards and real-time recommendations empower data-driven decision-making, driving business value and growth.
Benefits
Enhanced Scalability: Adoption by 5,000+ users across 3 countries, scaling for all divisions within the company.
Reduced Deployment Time: From an SLA of ~8 months in 2020 to 6-8 weeks production deployment of ML Models in 2023, enabling faster time-to-insight.
Increased Conversion Rate: ML Ops models contributed to a conversion rate increase of over 40%.
Projected Sales Lift: Expected sales lift of ~$15M, ~+50% versus the baseline in 2 quarters, with a 15% YoY growth in operating benefit.
Average Volumetrics: Over 350 TB of data loaded to Big Query/Cloud Storage, with ~100 TB data analyzed by ML models monthly across LOBs.
