
RAG Pipeline: 7 Iterations Explained!
Cyril Imhof
Bhavnish Walia, Senior Risk Manager AI/ML at Amazon, presents "Transforming Seller Onboarding in Retail: Responsible AI, RAG, and Risk Management." Timestamps: 00:00 Intro & overview of RAG, LLMs, responsible AI 00:30 Why document automation matters across industries 03:20 Manual onboarding workflow & motivation for automation 06:40 Challenges: human error, regulation, scalability 08:30 Technology evolution 2010 → 2025 10:30 Architecture – data pre-processing, layout models, RAG + LLM 13:30 RAG framework examples and retrieval logic 16:50 Explainability and human-in-the-loop learning 19:45 Evaluation metrics and impact (KPIs, efficiency gains) Onboarding new sellers onto retail platforms like Walmart and Amazon involves a complex, multi-step process designed to mitigate fraud and ensure regulatory compliance. One of the most critical steps is KYC verification, which traditionally requires manual review of identity, business registration, and compliance documents. This often results in long approval times and operational bottlenecks. To address these challenges, we leveraged foundational models with custom prompting, in-document summarization, and retrieval-augmented generation (RAG) powered by open-source LLM APIs. By automating document analysis and supporting human reviewers with AI outputs, we reduced onboarding time by more than 20%, improving seller experience and operational efficiency. Deploying AI in a regulated process like KYC required a strong responsible AI framework. We implemented guardrail models for edge case detection, anonymization protocols to protect sensitive data, privacy-preserving training techniques, and a rigorous validation pipeline to meet regulatory standards. This talk provides actionable insights for data scientists, compliance officers, regulators, and ML practitioners working at the intersection of AI, risk management, and regulation. Attendees will walk away with a practical framework for deploying AI in sensitive domains—covering risk strategies, scalable architectures, and lessons on balancing innovation with accountability. Official session recording from the Applied AI Summit 2025 Connect with us: Our website: https://www.johnsnowlabs.com/ LinkedIn: https://www.linkedin.com/company/johnsnowlabs Facebook: https://www.facebook.com/JohnSnowLabsInc X: https://x.com/JohnSnowLabs #ResponsibleAI #RAG #LLM #DocumentAutomation #RetailTech #AIinRetail #AIinFinance #RiskManagement #DigitalTransformation #SellerOnboarding