Thinking Fast With LLMs:Enabling Personalized Product Discovery & Recommendation

So, how long did it take you to choose the phone you have now?

Hours, weeks, months…. Well if you are on the farther end of the time spectrum you should attend this talk!

Some of us enjoy it in the name of retail therapy but for someone with a broken phone grappling with Nomophobia, it isn't fun. As e-commerce consumers, we face an eternal dilemma: finding the right product. Is there a way out of this maze?

However, this struggle extends beyond personal experiences. Software professionals building platforms encounter similar challenges, striving to attract user adoption and traction in a competitive market. Is there anything we can do more than better UX and a search bar? Maybe strap a recommendation engine to it?

Since the dawn of ChatGPT, LLMs are all the RAGe these days. But can they help us navigate the product jungle? So I set out to find how we can integrate LLMs to build effective product search and discovery and do context-aware personalized recommendations.

Do I need to know about LLM or Recommendation Engines to attend this talk?

• No, this talk is beginner-friendly.
• We will do a brief primer on these topics before we dive in.

Agenda

• Recommendation 101
• Motivation for Recommendation Engine
• Brief Overview of Use Cases
• Types of Recommendation
• High-level Overview of Classic Recommendation Algorithms
• Building a Recommendation Engine with LLM
• Implementing Classic Algorithms will LLMs
• Implementing a RAG pipeline for Recommendation

About the speaker

With over a decade of expertise in data wrangling and crafting data products, Prabakaran has developed a deep passion for leveraging technology to solve complex problems. Currently he is immersing himself in the realm of Platform Engineering.