Adapta Scales AI Education and Tools with Astra DB to Serve a Growing User Base

Adapta.org is a learning platform dedicated to empowering individuals and organizations to navigate the world of artificial intelligence (AI) confidently and easily. The company is recognized for its robust AI education and tool development offerings, emphasizing user-friendly solutions for complex AI tasks. Adapta.org offers a comprehensive range of educational resources, tools, and courses focused on practical, hands-on learning experiences that equip users with the skills to understand and apply AI in real-world scenarios.

Adapta

Products & Services

DataStax Astra DB

Industry

Education

Location

São Paulo, Brazil
Contact Sales

User Growth: Scaled to support 52,000 users.

Faster Search: Improved vector retrieval speeds.

Efficiency Gains: 35% faster development using Astra DB's SDK.

Challenges

As Adapta's user base grew to over 52,000 users, the company faced challenges in scaling its platform to meet increasing demands. The platform, which supports over 200,000 documents and 70,000 spreadsheets, relied on Pinecone as its vector database provider. However, Pinecone's size limitations restricted the amount of data Adapta could index, and there were performance constraints related to vector retrieval, which affected the speed of search queries.

Pinecone’s indexing and namespace management capabilities also presented challenges as the volume of documents and user queries increased, leading to inefficiencies. Additionally, the absence of local support made it difficult for Adapta to resolve technical issues and optimize its platform quickly. These challenges prompted Adapta to seek a more scalable, flexible, and better-supported alternative to meet their growing needs.

Solution

After evaluating multiple alternatives—including Qdrant and Weaviate—Adapta chose Astra DB. The decision was based on Astra DB’s superior scalability, faster vector retrieval, and better namespace management, which enabled Adapta to support its growing data needs. Astra DB’s local support team and flexible architecture gave Adapta the assurance it needed for ongoing expansion.

“We chose Astra DB because it removed the scaling limitations we faced with Pinecone. The speed of vector retrieval and the flexibility of Astra DB made a big difference for us. Additionally, having local support was crucial in ensuring a smooth transition and ongoing success,” said Lucas Braga, Head of Technology at Adapta.

Results

The migration from Pinecone to Astra DB was completed in just five days, a remarkably short timeframe considering the scale of the data involved. By leveraging Astra DB’s capabilities, Adapta now provides faster and more complex query support, addressing the needs of its diverse user base, including legal and healthcare professionals who use the platform for tasks like complex legal document analysis and medical imaging review.

Adapta has integrated LangChain into its platform to enhance document processing and vectorization, which is critical to its retrieval-augmented generation (RAG) capabilities. LangChain enables Adapta to efficiently handle large and complex documents, such as PDFs, Word files, and spreadsheets, by breaking them down into smaller, manageable chunks that are then vectorized to improve search and retrieval accuracy. This integration enables Adapta to deliver more precise and relevant results, even with extensive and intricate datasets. Additionally, Adapta is testing LangChain's ability to manage a growing volume of spreadsheet data, ensuring that large spreadsheets are effectively chunked, vectorized, and indexed to optimize search performance. This capability is essential as user queries become more complex and data types expand.

"With Astra DB, we've achieved faster development of RAG features,” said Braga. “The developer experience is outstanding—it's straightforward to use with clear documentation, and the level of abstraction in their SDK is impressive. My team gained about 35% of their time by using the Astra DB SDK.”

Looking ahead, Adapta is exploring the integration of DataStax’s Langflow AI PaaS. This no-code solution will enable them to rapidly prototype and validate new AI solutions and work seamlessly with Astra DB. This tool is expected to significantly speed up the development of minimum viable products (MVPs), enabling the company to respond more quickly to emerging customer needs. By combining the robust capabilities of LangChain with the flexibility of Langflow, Adapta aims to continuously improve its platform’s scalability and adaptability in the evolving AI landscape.

Migrating to Astra DB allowed Adapta to overcome significant scalability and performance challenges while delivering enhanced customer reliability. The partnership has enabled Adapta to focus on innovation, expanding its offerings without worrying about backend limitations. The close collaboration with DataStax has been crucial to Adapta’s ongoing success. As Adapta continues to grow and explore new opportunities, its partnership with DataStax remains a crucial factor in enabling it to stay agile and ready to meet the needs of its expanding user base.