The Challenge
Walmart needed an AI-powered product discovery experience that could handle millions of SKUs, understand natural language queries, and drive measurable engagement on product pages.
AI
Firework AI Platform
Spearheaded an AI-powered short video generation & live-streaming commerce platform leveraging LLMs, RAG, GraphRAG, and LangChain. Built AI chatbot for Walmart that increased product engagement by 400%.
Impact 400% ↑ engagement
0% Engagement Lift
0M+ Daily Queries
0ms Response Time
Case Study
My Approach
Built a multi-stage RAG pipeline using LangChain and LlamaIndex for document chunking, Azure OpenAI for embeddings and generation, and GraphRAG for relationship-aware retrieval. Deployed as a real-time chatbot on Walmart product pages with Kafka-based event streaming.
The Results
The AI chatbot increased product engagement by 400%, reduced support ticket volume, and became a key differentiator for Firework’s enterprise offering. The system handles millions of queries daily with sub-second response times.
Tech Stack
LangChainLlamaIndexAzure OpenAIRAGGraphRAGKafkaElasticsearchReactNode.jsGo