What is in this Course?

How to succeed in this course

Creating API keys and using them in VS code

Prompts and Tokens

Langchain

Dynamic Prompting

Different ways of Prompting LLM’s

Output parsers

Runnables

Chains in langchain

Memory in langchain

Assignment – 1

Vectors

Representing text as vectors

Text Similarity Search

Embedding models

What is RAG

Data Extraction

Text splitters

Selecting Embeddings

Testing for Bias

Selecting Vector Database

Qdrant Data Ingestion Code (Part-1)

Qdrant Data Ingestion Code (Part-2)

vanilla RAG pipeline

User Query Augmentation

Qdrant Search

RRF (Reciprocal Rank Fusion)

Reranking

Reranking (Code Walkthrough)

Selecting LLM

Creating Test Dataset

Metrics for retrieval

Metrics for LLM generation

Experiment tracking

Assignment-2

Introduction to LangGraph

Components of Langgraph

Conditional Edge

Parallel Processing

Dynamic Routing and Reducers

Chatbot with Memory

Chatbot execution

Callbacks

Introduction

React agent

Coding a React agent for planning

Compare Tool

Extract policy names

Extract policy data

Compare Tool Code

QA tool overview

QA tool code walkthrough

Testing

Assignment – 4

GenAI application lifecycle-1

GenAI application lifecycle-2

Technical Metrics to Monitor

Business metrics and iterative improvement

TCP, HTTP

HTTP versions

REST API principles

Client implementation

Server implementation

Testing HTTP API

Websocket

Disadvantages of Websocket

Code walkthrough of Websocket in local

Grpc protocol

Grpc Code walkthrough in local

Docker Overview

Container Orchestration

Compute and memory for Container Orchestration Tools

Overview

Pods and service in kubernetes

ELB, Deployment and HPA in kubernetes

VLLM

Components in server code

Need for Prometheus to expose semaphore metrics

Application code – gRPC setup

Application code – Traffic management and Streaming

Setting up AWS Infra using Terraform

Node Autoscaling using Karpenter

Deployment and HPA config

Service config

Deploying of Stability 3B model on AWS EKS G5 instance GPU’s

Load Testing using Locust

RAG pipeline overview

CI/CD Stages

Code build and Code Deploy

Setting up infrastructure for CI/CD Pipeline

Grpc for LLM call and Vector DB search

CI/CD pipeline for RAG using Github and Terraform

AWS Cloudwatch for monitoring metrics and alerts

Logging and Testing the API

Training components of conversational Ai

What are Neural Networks?

Training Neural Network

Classification Vs Regression Problems?

Cross Entropy loss

Similarity measures

Encoder only models

Decoder only models

Encoder-Decoder models

Training Language models

Need for GPU’s

Parameters of language models

Need for Quantization

Why we need Reinforcement learning to build advanced conversational Ai?

Reward Model

Reinforcement learning through Human Feedback(RLHF)

Capabilities and limitations of LLM’s?

Open source Vs Paid LLM’s in different scenario’s