Upgrade to Pro — share decks privately, control downloads, hide ads and more …

Gemini & Vertex AI: Revolutionizing Search with...

Avatar for Punsiri Boonyakiat Punsiri Boonyakiat
March 09, 2025
30

Gemini & Vertex AI: Revolutionizing Search with Generative API

Avatar for Punsiri Boonyakiat

Punsiri Boonyakiat

March 09, 2025
Tweet

More Decks by Punsiri Boonyakiat

Transcript

  1. Introduction Punsiri Boonyakiat - Lead Data Engineer @ Central Food

    Retail Group - Speaker @ Women Tech Maker 2024 - Speaker @ LINE DEV CONFERENCE 2021, 2024 Linkedin: https://www.linkedin.com/in/punsiri-boonyakiat-13909786/
  2. Agenda 01 Search behavior in each generation 02 POV: AI

    assisted search tools 03 Generative AI Trend 2024 vs. 2025 05 Vertex AI Agent Builder 01 Vertex AI Search 02 Gemini Multimodal x Vertex AI Search 06 Key Takeaway
  3. Multimodal AI: Unleashing the power of context AI agents: The

    evolution from chatbots to multi-agent systems Assistive search: The next frontier for knowledge work AI-powered constituent experience: seamless and personalized Security gets tighter— and tougher— with AI
  4. Challenge: LLM has limited knowledge, causing hallucinations User LLM What

    was Google Cloud's revenue in 2021? Hmm I actually don't know. Just going to guess...
  5. Why grounding with LLMs? LLMs are great at generating content

    but they fall short in a number of ways: 1. Generated content can be fake (hallucinations). 2. Generated content can be stale (not based on up-to-date information). 3. Generated content has no citation to its source. 4. Generated content has no relation to your private data (based on public knowledge only).
  6. Solution: Grounding / Retrieval Augmented Generation (RAG) Use search engine

    to enhance LLM capability and reliability User LLM Search engine Docs Hey Search, find reports on Google Cloud's revenue! What was Google Cloud's revenue in 2021?
  7. Grounding in Gemini with Vertex AI Search I need advanced

    groundings! Want to try Grounding in Gemini?
  8. Building a grounding is not so easy... Text Chunking 01

    02 03 04 05... Query Expansion Hybrid Search Knowledge Graph Reranking etc... Why not just use Google Search? It has them all. I need these advanced groundings to find right information!
  9. Vertex AI Search in Agent Builder: Google Search for enterprise

    data Extraction & Summarization Collection (web, files, DBs, connectors, etc.) Parsing Search Chunking & Annotation Tokenizing & Embedding Token & Vector Indexing Collect Build Serve Vertex AI Search in Agent Builder
  10. Vertex AI Search - Website Vertex AI Search - Unstructured

    Data - PDF Vertex AI Search - Semi-structured Data - JSON Vertex AI Search + Gemini Image understanding Integration with LINE Messaging API DEMO: Vertex AI Search