Case Study

Reklaim: Cloud-Native DevOps Transformationfor AI-Powered Luxury Goods Authentication

INDUSTRY

Retail

Project Type

DevOps

BaSIS Services

DevOps

Technology Leveraged

Amazon Nova, Amazon SageMaker, Yolov8, Meta SAM 2, Amazon ECS, Amazon ECR, Amazon Aurora, AWS ElastiCache, Amazon OpenSearch Service, AWS S2, Cloudwatch, AWS CloudTrail, AWS X-Ray, AWS IAM

Reklaim: Cloud-Native DevOps Transformation for AI-Powered Luxury Goods Authentication

Problem / NEED

Reklaim operates an AI-powered authentication platform for high-value luxury goods, helping buyers and sellers verify items across premium brands. As demand grew, the company began to outgrow its original infrastructure. The platform relied heavily on a single AI provider that delivered only 70–80% accuracy, forcing the team to rely on time-consuming manual verification to ensure trust and pricing accuracy.

At the infrastructure level, configuration management was fragmented across environments, creating security risks and operational inconsistency. As Reklaim expanded internationally and processed higher volumes, especially during major auction events, their existing setup struggled to scale reliably. The business needed a more resilient DevOps foundation that could support complex, multi-model AI pipelines, absorb sudden traffic spikes, and deploy updates without downtime in an environment where reliability directly impacts transaction value and customer confidence.

Solution

Basis partnered with Reklaim to modernize its infrastructure through a full cloud-native DevOps transformation on AWS. The platform was migrated from Elastic Beanstalk to a multi-cluster Amazon ECS architecture, with all infrastructure defined and managed through Infrastructure-as-Code.

The solution introduced five purpose-built ECS clusters: production backend, QA testing, Sidekiq workers, scheduled jobs, and web crawling services. Each is independently scalable to match its workload. GitHub Actions was implemented to deliver fully automated CI/CD pipelines, enabling frequent, zero-downtime production releases without service disruption.

At the AI layer, Reklaim evolved from a single-model dependency to a robust multi-model pipeline. AWS Bedrock was integrated alongside custom-trained models to improve reasoning, classification accuracy, and resilience. Security credentials were centralized using AWS Secrets Manager, improving security across environments.

This architecture gave Reklaim the flexibility to scale confidently while continuously improving model performance.

RESULTS

The move to a DevOps-enabled AWS platform delivered immediate and measurable impact. Reklaim’s authentication accuracy increased to 95%, up from 70–80%, directly strengthening pricing precision and customer trust.

The platform now supports thousands of concurrent users without performance degradation, even during peak auction periods. Processes that previously required days of manual effort are now completed in hours through automation. Replacing spreadsheet-driven workflows with integrated systems also improved visibility, reduced human error, and simplified day-to-day operations. Deployments are now predictable, low-risk, and frequent, with minimal downtime—dramatically improving overall platform reliability.

Other Key Outcomes

99.95% system uptime with sub-second response times, supporting real-time authentication workflows
Expanded coverage across major luxury brands and hundreds of bag and watch models
Reduced operational overhead by eliminating manual validation at scale
Cost efficiency and elasticity achieved through AWS-native services and managed GenAI infrastructure

Together, these improvements positioned Reklaim as a technology leader in luxury goods authentication, with a platform built for accuracy, speed, and trust at scale.