Arpit Agrawal (Elastiq.AI)
Arpit Agrawal (Elastiq.AI)
  • Видео 21
  • Просмотров 91 235
Why RAG is broken - Production Challenges for Enterprises while implementing AI.
Generative AI is solving a lot of the challenges for enterprises.
Everyone wants to implement Retrieval-Augmented Generation (RAG) and Agents. But RAG by design, is broken. Here are the real customer use-cases that show that show why RAG is not enough.
Arpit Agrawal, CEO of Elastiq, a Google Cloud AI-focused partner talks about the production architecture for implementing AI for enterprises.
#ai #genai #googlecloud #generativeai
0:00 Introduction
1:03 Google Agent Builder
2:51 Production Challenges
6:20 Challenge 1 - Data Aggregation
7:36 Challenge 2 - Negative Question
8:56 Challenge 3 - Tables and columns in documents
10:52 Challenge 4 - Infographics in documents
12:35 Challenge 5 - Geospatial...
Просмотров: 328

Видео

Search Videos and Images with AI
Просмотров 2615 месяцев назад
Use Generative AI to search through hours of video in an instant. No need to manually tag metadata. Now use AI and search within the semantic meaning of the video and go directly to the point of interest in the video.
Data Privacy with GenAI
Просмотров 317Год назад
Design for data privacy when working with Generative AI
Generative AI will disrupt Customer Service, here's how - 20 use-cases.
Просмотров 2,9 тыс.Год назад
00:00 Description 4:21 Government Regulation 7:01 Impact on Customer Service 10:51 Virtual Agents 18:32 Multi Channel 21:53 Multi Modal 24:20 Personalization 29:58 Caller Intent Prediction 32:20 Speaker ID Auth 35:54 Brand Voice - custom speech voice 40:06 Compassionate Virtual Agent 42:54 Sentiment Analytics 46:30 Multi Lingual Support 48:21 24*7 Support 50:31 Agent Assist 52:07 Faster Trainin...
Generative AI for Insights Discovery and Knowledge Management
Просмотров 767Год назад
Using Generative AI to get insights from data.
Nova - AI for Insights Discovery
Просмотров 508Год назад
AI for Insights discovery! NOVA can connect to any data store and help you surface insights from data, be it a data lake, a relational or a non-relational database, just a bunch of documents in a drive, or even a 3rd party API. Reduce time to insights. Intuitively query any data store in natural language. Democratize access to data. Anyone can now be a data analyst!
Solved: How to Build the Perfect Analytics Team
Просмотров 4562 года назад
GCP.LIVE Podcast EP - 001 The Roles and Tools for Analytics with Arpit Agrawal and Aniket Goel Arpit Agrawal: www.linkedin.com/in/arpitdotme/ Aniket Goel: www.linkedin.com/in/aniketgoel
Machine Learning with SQL - BigQuery ML
Просмотров 3,2 тыс.3 года назад
Now you can train Machine Learning models using SQL!! 0:00 Introduction 0:24 The use case 2:13 The data 3:33 Model Training 6:46 Batch Prediction 8:23 Real-time Prediction 12:17 Conclusion Read more here: gcp.live/machine-learning-with-sql-bigqueryml-vertex-ai/
Querying 100 Billion Rows using SQL, 7 TB in a single table
Просмотров 52 тыс.3 года назад
0:00 Introduction 0:59 The data 1:22 1K row Query 3:53 100K row Query 4:32 10M row Query 5:20 1B row Query 6:04 100B row Query 8:03 Query Costs 8:45 Conclusion
Data Journey EP-02: Batch Ingestion 📦 - 5 ways to ingest files into Google Cloud
Просмотров 13 тыс.4 года назад
Data Journey EP-02 Batch Ingestion - 5 ways to ingest files into Google Cloud Intro: (00:00) Data Ingestion Challenges: (00:39) Data Ingestion Design: (01:32) Ingestion Option-1 : Cloud Console: (05:42) Ingestion Option-2 : GSUTIL: (06:17) Ingestion Option-3 : Storage Transfer Service: (07:45) Ingestion Option-4 : BigQuery Transfer Service: (08:55) Ingestion Option-5 : Storage Transfer Applianc...
Data Journey EP-01: Introduction 📈 - What to expect and Why Google Cloud?
Просмотров 1 тыс.4 года назад
Data Journey EP-01: Introduction 📈 - What to expect and Why Google Cloud?
Google Cloud Networking & Security for the Enterprise - Creating a custom VPC
Просмотров 3974 года назад
Google Cloud Networking & Security for the Enterprise - Creating a custom VPC
Rapid Response Virtual Agent for Contact Centers powered by Google Cloud
Просмотров 4044 года назад
Rapid Response Virtual Agent for Contact Centers powered by Google Cloud
Google Cloud Discounts - 7 secrets (+1 Bonus) that help you save Cloud Costs
Просмотров 8174 года назад
Google Cloud Discounts - 7 secrets ( 1 Bonus) that help you save Cloud Costs
Google Colaboratory (Colab) vs Google Cloud AI Platform Notebooks
Просмотров 6 тыс.4 года назад
Google Colaboratory (Colab) vs Google Cloud AI Platform Notebooks
AI Platform Notebooks - Managed Jupyter Notebooks Google Cloud
Просмотров 5 тыс.4 года назад
AI Platform Notebooks - Managed Jupyter Notebooks Google Cloud
Why Google Cloud for Analytics & AI?
Просмотров 6704 года назад
Why Google Cloud for Analytics & AI?

Комментарии

  • @JunaidKhan-gq8nw
    @JunaidKhan-gq8nw Месяц назад

    Great, Thanks a lot, sir.

  • @prasannatakumargautam221
    @prasannatakumargautam221 2 месяца назад

    Hi sir . I have one query regarding big query

  • @abhijayrajvansh
    @abhijayrajvansh 3 месяца назад

    it's always an Indian guy!

  • @the_shree
    @the_shree 3 месяца назад

    Clear explanation 😊

  • @swapnali_20
    @swapnali_20 4 месяца назад

    Very well explained 👏

  • @houssem25000
    @houssem25000 5 месяцев назад

    So I don't have to carry about performance when I make projects ?!

  • @loufoua7640
    @loufoua7640 5 месяцев назад

    thank you very much very well explained

  • @yolemmein
    @yolemmein 5 месяцев назад

    Very informative. Thanks.

  • @vikasbohra3253
    @vikasbohra3253 6 месяцев назад

    Very nice 👌🏻👌🏻👌🏻👌🏻👏🏻👏🏻👏🏻

  • @mulshiwaters5312
    @mulshiwaters5312 6 месяцев назад

    any mechanism to move files from GCP (GCS Storage) to another Cloud Provider like Azure ?

  • @Mju98
    @Mju98 8 месяцев назад

    Hello sir. I tried to import 400k data into big query sandbox. But ended with more errors. Is this possible to import those data. Pls anyone help me it's urgent ( interview assignment)

  • @Rpskmr
    @Rpskmr 8 месяцев назад

    Nice video but while voicing better to expand the screen than side by side videos

  • @vijaysoni6517
    @vijaysoni6517 9 месяцев назад

    Amazing 👏

  • @samuelbarros4005
    @samuelbarros4005 10 месяцев назад

    and how can i do that to predict how much sales really have using sales and refunds?

  • @Hrzzz1
    @Hrzzz1 Год назад

    we can download this database to do some testes ? I nice ideal for next video is compare this same situation with noSQL database.

  • @jiyasingh664
    @jiyasingh664 Год назад

    Amazing explanation for beginners, thanks a lot for this informative video!

  • @vijaybhaskar5333
    @vijaybhaskar5333 Год назад

    Good discussion. Informative. you guys deserve more views.

  • @FlowGPTOfficial
    @FlowGPTOfficial Год назад

    Visit www.flowgpt.com and join the world's largest prompt community!

  • @DanThe1Man
    @DanThe1Man Год назад

    Nice video. I see this asked a lot, too.

  • @DanThe1Man
    @DanThe1Man Год назад

    Great talk.

  • @nfacundot
    @nfacundot Год назад

    Hello, can I connect it on php?

  • @davidlean8674
    @davidlean8674 Год назад

    This is nice but not that impressive. Obviously, the table is being stored using Columnstore Compression techniques. So you only need to query the columns in the select list. And they are typically grouped in blocks of 1 M or more. These header pages keep rowcount values. So you are not reading every row. Just the block headers of a single column. If your query forced the scan of all rows in the "block" asking it to be combined with other fields in the same row or in other tables before you could filter it. You will no longer be in the columnstore sweet spot. and the difference in query speed would be more striking. Still good thou, as that is a common use case.

  • @ayanroy08
    @ayanroy08 Год назад

    Very well explained

  • @swapnali_20
    @swapnali_20 Год назад

    Data privacy is an important concern when it comes to interacting with AI systems. Nicely explained 🎉

  • @baldevsingh-b5h9j
    @baldevsingh-b5h9j Год назад

    awesome content..thanks for such videos

  • @toxiclife
    @toxiclife Год назад

    what to do when I want to overwrite 100 millions of rows into new table, in minutes? df.write.mode("overwrite").saveAsTable("FINAL"), if you could please help with this?

  • @elastiqai
    @elastiqai Год назад

    Watch the full podcast here: ruclips.net/video/UfBRIEcRAmg/видео.html

  • @swapnali_20
    @swapnali_20 Год назад

    Great👍

  • @TheElementFive
    @TheElementFive Год назад

    The first question you should always ask when working with a 100 billion row database: “Why do I have a 100 billion row database?”

    • @davidlean8674
      @davidlean8674 Год назад

      And the answer would be "because I work with a multinational enterprise customer". If you have a large market share in China (1 bill people) , India (1 Bill people), Europe 0.75 Bill, USA (350M people) it doesn't take long to get to 100 BIllion transactions. If you want to do Financial Year on Year comparisons, you need to keep at least 24 months of data, usually 36 months. .

  • @mathteacher5670
    @mathteacher5670 Год назад

    excellent sir thank you so much highly motivational for passionate person

  • @gbohra
    @gbohra Год назад

    Guys, this is clear and concise as possible, while still providing enough information to be helpful.

  • @JaiGurujishukranaGuruji947
    @JaiGurujishukranaGuruji947 Год назад

    Excellent

  • @YashPatil-pm2xc
    @YashPatil-pm2xc Год назад

    Very cool

  • @abhijeet6785
    @abhijeet6785 Год назад

    This is really very informative!!

  • @vishnuiyengar8555
    @vishnuiyengar8555 Год назад

    Great video. Thank you so much for such a clear and clean explanation. One question I had was that I am planning to use a notebook environment for my senior design project during college. Our group is planning to use this for implementing GAN and YOLO models to detect defects in 3D printing. Since we are planning to use at least 1000 images for the training dataset for each type of defect, we wanted to know which notebook is good. Do you think Google Colab will be a good fit for such a project?

  • @muzahmad2104
    @muzahmad2104 Год назад

    Great content keep it up. But sort the echo please.

  • @nileshmchavda
    @nileshmchavda Год назад

    fantastic!

  • @ChuckWilliams-gt2yb
    @ChuckWilliams-gt2yb Год назад

    Very, very nice!!!

  • @ShivansuSoni
    @ShivansuSoni Год назад

    Amazing

  • @vikasbohra3253
    @vikasbohra3253 Год назад

    Very innovative. Thanks for explaining it in a very precise manner. 👌🏻👌🏻👍🏻👍🏻

  • @MDDM03
    @MDDM03 Год назад

    marketer of google cloud.. nothing states what to improve

  • @appstuff6565
    @appstuff6565 Год назад

    Very well Explained. Thanks!

  • @dattatreyakomakula1650
    @dattatreyakomakula1650 Год назад

    Excellent sir.

  • @kaijunhe3339
    @kaijunhe3339 2 года назад

    Great Video, super helpful

  • @christinamariehicks1078
    @christinamariehicks1078 2 года назад

    Only one lady im 65 years old my name christina smith hicks ..

  • @karan7843
    @karan7843 2 года назад

    Brilliant content. Would love if more videos' on same series would come often.

  • @aminremiiii
    @aminremiiii 2 года назад

    Please for 50 days I am looking for this i wanna to create 2000 users in mysql and set the phone number as user name and password my be say me how can i create most users with default password? That's

  • @ilducedimas
    @ilducedimas 2 года назад

    Thank you my good sir.

  • @muhamadridwan4766
    @muhamadridwan4766 2 года назад

    wow!

  • @arthurrodrigues5382
    @arthurrodrigues5382 2 года назад

    Amazing!