Veriff is an identity verification platform partner for organizations in various industries. They use advanced technology, including AI-powered automation and human feedback, to verify user identities. Veriff standardized their model deployment workflow using Amazon SageMaker, reducing costs and development time. They use SageMaker multi-model endpoints and Triton Inference Server to manage and deploy ML models efficiently. This solution has led to a significant cost reduction and faster model deployment for Veriff.
Veriff: Streamlining Model Deployment with Amazon SageMaker
Veriff is an identity verification platform trusted by leading organizations in finance, gaming, and more. They combine AI-powered automation with human expertise to ensure trust in user identities throughout the customer journey.
Infrastructure and Development Challenges
Veriff faced challenges in deploying and managing their machine learning (ML) models, which ranged from lightweight to complex computer vision models. Their existing solution required manual provisioning of GPU instances and building REST API wrappers for each model, resulting in operational overhead and suboptimal cost profiles.
Solution Overview
To address these challenges, Veriff adopted Amazon SageMaker’s multi-model endpoints (MMEs) and NVIDIA’s Triton Inference Server. MMEs allowed them to deploy and manage a large number of models efficiently, reducing hosting costs and deployment overhead. Triton Inference Server simplified the process of building REST APIs from models and enabled the deployment of model ensembles.
Model Versioning and Continuous Deployment
Veriff implemented a monorepo approach for managing their models, using Pants for code management and applying code quality tools and unit tests. They integrated this monorepo with a continuous integration (CI) tool to automate the deployment process, ensuring model quality and versioning.
Cost and Deployment Speed Benefits
By leveraging SageMaker MMEs, Veriff reduced model development time from 10 days to an average of 2 days. They also achieved a 75% cost reduction in GPU model serving compared to their previous Kubernetes-based solution. The auto scaling features of SageMaker allowed them to optimize costs based on traffic patterns.
Conclusion
Veriff’s adoption of Amazon SageMaker MMEs streamlined their model deployment workflow, reducing costs, improving efficiency, and maintaining performance. Their CI/CD pipeline and model versioning mechanism serve as a reference implementation for combining software development best practices with SageMaker MMEs.
Unlock the Power of AI for Your Business
Discover how AI can revolutionize your company and stay competitive. Identify automation opportunities, define measurable KPIs, select the right AI solution, and implement gradually. For AI KPI management advice, contact us at hello@itinai.com. Stay updated on AI insights through our Telegram channel t.me/itinainews and Twitter @itinaicom.
Spotlight on a Practical AI Solution: AI Sales Bot
Explore itinai.com/aisalesbot, an AI-powered sales bot designed to automate customer engagement and manage interactions throughout the customer journey. Discover how AI can redefine your sales processes and customer engagement.
