Get RAG Into Production in 15 Minutes | Rajiv Shah, Contextual AI

Get RAG Into Production in 15 Minutes | Rajiv Shah, Contextual AI

by Rajiv Shah

Applied Sciences & Engineering AI — Models & Research UYQArtificial intelligence (AI)UYQFGenerative artificial intelligence
🎙 Rajiv Shah 👥 5K 📅 September 29, 2025 ⏱ 23 min 👁 74 📄 tutorial 🧭 2026-08-15
Available in: English (current) Français

Keywords

RAGretrieval-augmented generationMCPproductionmanaged service

Summary

In this session from GenAI World, Rajiv Shah, Chief Evangelist at Contextual AI, demonstrates how to build and deploy an end-to-end RAG pipeline in 15 minutes. He emphasizes treating RAG as a managed service, similar to how one would use a vector database or foundation model API. The talk covers the key components of a production-ready RAG system: data understanding (parsing, OCR, vision-language models), hybrid retrieval (semantic and lexical), reranking, and grounded generation. Shah shows a live demo of creating an agent, connecting it to a data store, and querying it, highlighting features like content inspection, feedback mechanisms, and adjustable knobs. He also discusses advanced topics such as query reformulation, metadata filtering, and using RAG within agentic workflows via MCP. The presentation concludes with practical advice on getting started with Contextual AI’s platform, which offers a consumption-based pricing model and free credits. The talk is aimed at developers and practitioners looking to move RAG from prototype to production efficiently.

159 words

Critical Evaluation

Value of the Information & Strength of the Argument

The talk provides valuable insights into the practical challenges of production RAG, such as data extraction complexity and the importance of reranking. Shah argues convincingly for a managed service approach, citing the burden of maintaining custom pipelines. He supports his claims with references to research showing productivity gains from RAG and demonstrates the platform’s capabilities. However, the argumentation is largely based on anecdotal experience and product features rather than rigorous comparative analysis.

Scientific Rigor, Source Quality, Title Accuracy

The scientific rigor is moderate: the speaker references a recent research paper on RAG’s impact but does not provide specific citations. The sources cited are limited to the MLOps World link in the description. The title accurately reflects the content, which is a tutorial-style demonstration. The talk is more of a product pitch than a neutral educational piece, but it does include useful best practices.

152 words

Title / Content Match

The title accurately reflects the content: a hands-on session demonstrating how to build and deploy a RAG pipeline in 15 minutes.

Quality & Reliability

7/10

The speaker is a practitioner with industry experience, and the content is based on practical implementation. However, it is primarily a product demonstration with limited independent verification of claims.

Key Moments

Markers derived by PSI from the transcript: the creator did not define chapters.

Cited Sources

  • MLOps World — Mentioned as the event organizer and source for more information.

Concurring Sources

Contribution & Novelties

The talk provides a practical, hands-on perspective on deploying RAG in production, emphasizing the value of managed services. It highlights specific techniques like hybrid retrieval, reranking, and metadata filtering that are often overlooked in introductory material. The demonstration of MCP integration for agentic workflows is a notable addition.

Pour aller plus loin :

92 words

Radar Profile

The radar shows a balanced profile with moderate scores across all dimensions, indicating a solid but not exceptional presentation. The highest scores are in information quantity and quality, reflecting the practical content, while technical depth and reliability are slightly lower due to the product-focused nature.

Reliability 6/10

💬 No comments were provided for analysis.