วีดีโอ

Brilliant. This video should be played in higher schools and colleges (technical and management). so much to learn these Amagi guys. simply wow. Wish this video gets promoted more for betterment of society.

Great!

Everyone is indian why speaking to english

Great job 💯👏 sud n summit bhaiya 🎉 congratulations🥳

You ppl are doing great jobs ,keep it up guyzzz

Is Mr BAGRI still in NIYO? I am an NRI NIYO customer and have SIP through them. Apparently it seems his company and he himself have gone underground.. I suspect being defrauded and now thinking to make a formal complaint SEBI and alert other users via various Social media platforms

unethical company; revoked my friend's and the entire selected batch's offer from IIT Guwahati at the last moment, last year ! The placement season was over, and thus, wasted my friend's year !! They are blacklisted from IITs placements. This shows their ethics !

That’s why Raghu is called ‘The iron man’, insightful master class. Beautiful portray of supply side issues and their resolutions.

Bro right now they are focussing on privalege and tier 1 consumer but once they reach scale they will start entering into smaller cities and play price game there

❤❤❤❤❤

Good content but the format of text question and audio answer is really irritating. This seems to be a podcast and we watch it while working or doing some chores, but to repeatedly look at the screen for questions is frustrating ! Either be complete text format or complete podcast, don't mix up, making it difficult to consume ur content. Rest is good. 👍🏼

Hey Seed we don't Need Background Music THIS SPOILS the Whole Experience. Plz From Next Episode No Background Music

Stage ✌️👍🔥

"Great initiative by STAGE - OTT to promote our cultural diversity!"

Insightful 38 minutes

Great content, but very irritating fake laughs by the moderator

Baate badi badi karte he par sambhalna bhi nai aata business logo ko. Inke delivery log hamesha pizza ko kharab ho delivery karte he.

चोर कंपनी है swiggy इन दोनो gado ने मिल के भारत के लड़को या युवा को मजदूरी करने के लिए गुलाम बनाने का tech support liya hai.= Mdrcd

thank you for this

Best tutorial on the topic !!!

Just to put the chat into context - “Blackbuck posted a Rs 290 crore loss in FY23 as compared to Rs 285 crore in FY22. Its ROCE and EBITDA margins stood at -76% and -36% respectively. On a unit level, it spent Rs 1.44 to earn a rupee in FY23.”

13:03 : we should have a system where a guest should disclaimer on interest. With 700 cr losses and rev of 100 crores should such cash burn be a role model for listeners ?

My heart girish anna

❤❤

inspiring to hear Thiru's journey and challenges - and the Accel team also is so cheerful and down-to-earth... its heartening and encouraging..

Please ..?

❤

Great. video. Literally learned a llot. Creating a pitch deck on my own now. even nows its hard to researc about finance, and lot. But girish sir made pitch deck solo those days. hats off.

Curious why Accel invests in marketplaces when these are inherently hard and have a slower growth rate?

Thank you for doing this!

I wish this was even longer. So good!

Girish Sir ... Thank you for being my mentor via youtube

Such a wonderful video 🎉. Happy to see our Girish and Shekhar in this video. Lovely one.

auto insaurance is a hyper locallized bussiness no customer underwriting in old model no interecation points , people look insaurance companies as someone who will pay their bills , while they should be larger

Can you please share the document? It’d be really really helpful

0:00:00 - Introduction 0:01:55 - Overview of Excel Atoms Accelerator Program 0:03:28 - How to Apply to Excel Atoms 0:04:39 - Serverless GPUs - Introduction 0:05:18 - Why Custom Models? 0:08:07 - Challenges in Machine Learning Deployments 00:22:56 - Kubernetes and Scaling Challenges 00:23:17 - Avoiding Kubernetes for Generative AI Models 00:23:36 - Service Level Costs in Kubernetes 00:23:59 - Serverless for Unpredictable Usage 00:24:18 - Benefits of Serverless for Spiky Loads 00:24:38 - Factors Affecting Scaling in Kubernetes 00:24:58 - Serverless for Large Model Sizes 0:45:26 - Building the Platform 0:46:02 - Unique Problems Solved 0:48:13 - Demo of the Product 0:52:01 - Who Can Use Infoless? 0:53:15 - Comparison with Model.com 0:55:17 - Inference Stacks for Different Use Cases

0:00:39 - Purpose of the Webinar Series 0:02:19 - Introduction to True Lens 0:03:31 - Challenges with Language Models (LLMs) 0:06:10 - Addressing Hallucinations with RAGs 0:08:26 - Evaluating RAG Models with True Lens 0:10:34 - Use Cases and Applications 0:14:44 - Debugging Hallucinations with the Triad 0:33:43 - Questions and Logging 0:34:04 - Leaderboard and Feedback Functions 0:34:35 - Trying Different Distance Methods 0:35:58 - Systematic Experimentation and Evaluation Pitfalls 0:37:13 - Evaluating Application Using TrueLens 0:38:04 - Failure Modes and Comprehensive Evaluation 0:39:20 - Data Agents and External Knowledge 0:40:25 - Chatting About Agents Application

0:00:30 - Future of LLMs 0:01:22 - Speaker's Background 0:02:20 - Current State of LLMS 0:03:49 - Gorilla: Connecting LLMs with APIs 0:22:44 - API Calls and Scaling 0:23:07 - Scaling with Proper Tools 0:23:27 - Embracing Parallelism 0:23:46 - Comparison between Zero-shot and Retrieval Augmented Models 0:24:09 - Data Generation and Inference Pipeline 0:24:34 - Time Check 0:24:54 - User Prompt and Reference API Selection 0:25:14 - Zero-shot Approach 0:25:48 - Evaluation of Hallucination 0:45:15 - Using APIs and Building Composable APIs 0:48:04 - Deprecation of APIs and Model Adaptation 0:50:43 - Handling Layers and Evaluating AST

0:00:09 - Introduction to the Webinar 0:00:56 - Introduction to LangChain 0:02:41 - Overview of Building Blocks 0:05:06 - Overview of Language Models 0:08:28 - Use Cases: Retrieval and Augmented Generation 0:09:23 - Use Cases: Summarization 0:25:17 - Introduction to Langsmith 0:25:39 - Usefulness of Langsmith as a Tooling Framework 0:26:04 - Collecting Feedback with Langsmith 0:26:22 - Editing Output within Langsmith 0:26:42 - Creating Ground Truth Datasets 0:27:05 - Evaluating App Performance with Langsmith 0:50:29 - Data Differentiation in ML 0:51:08 - Addressing Hallucinations in Medical Data 0:53:05 - Conclusion and AI Accelerator Program

0:00:11 - Introduction to Generative AI 0:01:18 - Evolution of Generative AI 0:03:05 - Attention Mechanism in Language Models 0:04:24 - Encoder-only and Decoder-only Models 0:06:10 - Variational Autoencoders (VAEs) 0:10:02 - Few-Shot Learning and Prompt Engineering 0:12:50 - QnA 00:19:53 - Model Fine Tuning vs Prompt Tuning 00:20:09 - Model Fine Tuning Approach 00:21:31 - Use Cases and Trade-offs in Fine-Tuning 00:24:25 - Prompt Tuning and its Benefits 0:40:26 - Which ARM core architecture is used in Sitara? 0:40:49 - Retrieval augmentation for generating responses 0:41:11 - Accuracy in generative AI for industries like Financial Services 0:42:18 - Accuracy evaluation and considerations in generative AI 0:43:08 - Use cases requiring high accuracy and traceability 0:43:37 - Power of generative capability combined with embeddings 01:01:01 - Introduction to Bedrock and API Service 01:01:16 - Secure Data Deployment with Bedrock 01:01:32 - Value Proposition of Bedrock 01:02:03 - Wide Range of Compute Instances in AWS 01:02:41 - AI Stack Options in AWS 01:03:15 - SageMaker JumpStart and Fine-Tuning Models 01:03:45 - Availability of Publicly and Proprietary Models

00:00:11 - Introduction to the State of Image and Video Generation 00:01:29 - Challenges in Deploying Generative Models 00:02:08 - Benefits of Using Segment APIs 00:03:15 - Companies Acquiring Millions of API Requests 00:04:35 - Introduction to Generative Models 00:05:26 - Workflows in Image and Video Generation 00:24:06 - Prompting and Fine-Tuning Stable Diffusion 00:25:33 - DreamBooth Technique 00:26:36 - Lora Technique 00:27:53 - Advantages of Lora 00:29:35 - Merging Styles with Lora 0:49:45 - Leveraging Models at Scale and Understanding Use Cases 0:50:05 - Real-Time VR/XR Rendering Engines 0:50:22 - Detecting Fake Content 0:51:41 - Fenaki Library for Longer Videos 0:53:18 - Workflow Examples - Storyboarding and Fashion 0:55:56 - Scribble Model Demo

00:00:13 - Introduction to Llama Index 00:01:36 - Overview of Llama Index Framework 00:04:49 - Components of Llama Index Framework 00:08:40 - Use Cases of Llama Index 00:16:25 - Building a Query Interface Engine 00:19:35 - Handling Multi-step Queries 0:28:53 - Demo of Loading a Book 0:29:00 - Accessing Different Chapters in the Book 0:29:08 - Options for Querying Information 0:29:33 - Retrieving Information about Chapters and Characters 0:30:04 - Questions on Masking and Accuracy 0:30:32 - Using Open Source Models for Masking 0:30:55 - Training a Model for Masking 0:31:23 - Limit 0:59:22 - How does the router retrieve information from a collection of books? 1:00:02 - Filtering notes related to books 1:01:07 - Classifying queries and retrieving relevant information 1:02:16 - Using L-index with private data 1:03:22 - Conclusion and closing remarks

0:00:00 - Introduction 0:01:30 - Goals of the Webinar Series 0:02:26 - Introduction to Speakers 0:04:17 - Introduction to Hashnode 0:05:16 - Introduction to Rigs - AI Companion 0:06:01 - Launch of Rigs 0:28:00 - The Safest and Fastest Models 0:28:24 - Speed and Data Retrieval 0:28:47 - Summarizing Search Results 0:29:07 - Cost-effective and Code Generation 0:29:29 - Long Context History 0:29:49 - Choosing Between Versions 0:30:33 - Better Reliability and Rate Limiting 0:31:16 - Request 0:54:54 - Query Caching and Similarity 0:55:11 - Development Process for Building a Feature 0:56:24 - Community Involvement in Improving Functionality

0:00:11 - Starting the Webinar 0:00:31 - Introduction to AI Chatbot Evolution 0:01:40 - Different Perspectives on Building Chatbots 0:02:22 - Introduction to Axel and AI Investments 0:03:27 - Announcing AI Cohort Program 0:04:41 - Building a Support Bot with Lang Chain 0:09:48 - Code Walkthrough for Support Bot 0:19:27 - Using Pinecone 0:28:55 - Dimensions for Open AI Embeddings 0:29:09 - Loading the Index 0:29:20 - Naming Convention and Metric Choice 0:29:41 - Splitting Documents and Providing Metadata 0:30:06 - Helper Functions for Conversational Retrieval 0:30:55 - Streaming Responses with Lang Chain 0:31:14 - Parameters for Conversational Retrieval 0:31:40 - Returning Source Documents 0:56:24 - Single Step Tasks 0:56:42 - Complex Systems 0:57:16 - Limitations of Automation 0:57:48 - Logical Framework for Automation 0:58:25 - Examples of Logical Frameworks 1:00:09 - Evolution of Conversational AI 1:04:24 - Leveraging AI for Automation 1:18:24 - Task Manager and GPT System 1:18:43 - Pros of GPT Systems 1:19:22 - Advantages of AutoGPT 1:19:39 - Communication Challenges with Automation 1:20:17 - Line Chain vs. GPD in B2P World 1:20:34 - Cons of Lang Chain and AutoGPT 1:22:05 - Handling Training Data for GPT 1:40:03 - Taking Questions and Answers 1:41:03 - Using GPT and Hybrid Architectures 1:42:10 - When to Use GPT and Other Models 1:48:21 - Human Labeled Cross Source Data for Fine-tuning 1:55:59 - Logging Chatbot Conversations for Training Data 1:59:41 - Conclusion and Closing Remarks

0:00:11 - Starting the Webinar 0:00:31 - Introduction to AI Chatbot Evolution 0:01:40 - Different Perspectives on Building Chatbots 0:02:22 - Introduction to Axel and AI Investments 0:03:27 - Axel's AI Cohort Program 0:04:41 - Building a Support Bot with LangChain 0:09:48 - Code Walkthrough for Support Bot 0:19:27 - Using Pinecone for Vector Database 0:28:55 - Dimensions for Open AI Embeddings 0:29:09 - Loading the Index 0:29:20 - Naming Convention and Metric Choice 0:29:41 - Splitting Documents and Providing Metadata 0:30:06 - Helper Functions for Conversational Retrieval 0:30:55 - Streaming Responses with Lang Chain 0:31:14 - Parameters for Conversational Retrieval 0:31:40 - Source Documents and Return Source 0:56:24 - Single Step Tasks 0:56:42 - Complex Systems 0:57:16 - Limitations of Automation 0:57:48 - Logical Framework for Automation 0:58:25 - Examples of Logical Frameworks 1:00:09 - Evolution of Conversational AI 1:04:24 - Leveraging AI for Automation 1:18:24 - Task Manager and GPT System 1:18:43 - Pros of GPT Systems 1:19:22 - Advantages of AutoGPT 1:19:39 - Communication Challenges with Automation 1:20:17 - Line Chain vs. GPD in B2P World 1:20:34 - Cons of Lang Chain and AutoGPT 1:22:05 - Managing Training Data for GPT 1:40:03 - Taking Questions and Answers 1:41:03 - Using GPT and Hybrid Architectures 1:42:10 - When to Use GPT and Other Models 1:48:21 - Human Labeled Cross Source Data for Fine-Tuning 1:55:59 - Logging Chatbot Conversations for Training Data 1:59:41 - Conclusion and Closing Remarks

0:00:00 - Introduction 0:01:18 - Purpose of the Webinar 0:03:08 - Discussion Topics 0:05:07 - Survey Results 0:09:00 - Why VCs are Excited About AI 0:10:09 - Understanding the AI Ecosystem 0:13:12 - Different Models in the AI Landscape 0:15:16 - Challenges with Large Language Models (LLMs 0:25:01 - Introduction to Language Models 0:25:22 - Summarization and Sentiment Analysis 0:25:42 - Text Classification for Sentiment Analysis 0:26:00 - Question Answering with Language Models 0:26:21 - Understanding Questions for Question Answering 0:26:41 - Question Answering with OpenAI's ChatGPT 0:26:59 - Semantic Search in Question Answering 0:52:01 - Introduction to Prompt Design 0:53:16 - Using Prompt Templates 0:54:17 - Prompt Management 0:55:32 - Introduction to Lang Chain Framework 0:57:41 - Semantic Search with Vector Databases 1:00:45 - Maintaining Context in Conversations 1:02:16 - Introduction to CleverTap's AI-powered Product 1:18:46 - Development of Models and Comparison with GPT 1:20:11 - Understanding Model's Aim and Purpose 1:21:47 - Fine-Tuning Models for Personalization 1:23:22 - Fine-Tuning Models for Customer Language Lingo 1:45:27 - Experience at Excel and Learning New Things 1:45:49 - Everything AI Excel Group 1:46:55 - Recommended Channel for Beginners - How I Learned Lang Chain 1:47:42 - Question on Stable Diffusion Models 1:49:06 - Talk on Advanced Stable Diffusion Techniques

can you paste the link to notebook

'promosm'

Thank you Mr Vetri for sharing your insights.

Very interesting, keep it up