FuseBase: AWS migration with SharePoint data pipeline
“Working with Tech 42 was a seamless experience from start to finish. Tech 42 team demonstrated exceptional technical expertise and communication throughout the project, building a robust SharePoint data pipeline to feed our new data lake. The team was proactive, responsive, and always attentive to our needs — addressing feedback quickly and ensuring smooth collaboration through shared Slack channels and regular check-ins. The project exceeded expectations in quality and delivery, providing a solid foundation for future integrations. We look forward to continuing our partnership with Tech 42 as we expand our data connectivity capabilities.”
Denis Kibalko
CTO
Project summary
Fusebase's AI agents needed access to customer SharePoint repositories to provide contextual responses, but no self-service integration existed. Each connection required custom work, and the company lacked a foundation for multi-source connectivity. Tech 42 built a serverless data ingestion pipeline using Lambda, API Gateway, DynamoDB, S3, and EventBridge. The architecture enables self-service, multi-tenant SharePoint connections with isolated data storage for each customer. Secrets Manager secures credential management. EventBridge schedules and triggers pipeline execution. The work also supported Fusebase's migration from OpenAI to Amazon Bedrock for AI inference. The pipeline established the foundation for future data connectors. All infrastructure, APIs, and pipeline logic are documented and tested. Fusebase can now expand connectivity to additional data sources while maintaining security and tenant isolation.
