Tag: data-science

Python contd Variables Refer Here for variables # Variable a = 5 # type hinting a:int = 5 b = 5.0 # type hinting b:float = 5.0 c = "hello" # type hinting c:str = "hello" Operators Refer Here for operators Refer Here Python indentation Block: Block in python refers to one or more lines… Continue reading Gen-AI Developer Classroom notes 16/Sep/2025

Langgraph Refer Here for langgraph docs Langraph helps in building workflows with graph as a central core concept This is a framework built on Langchain that is used for creating stateful multi agent workflows using a graph. Graph is made up of nodes and edges Each node can be a tool, LLM, retriever or custom… Continue reading Gen-AI Developer Classroom notes 14/Sep/2025

Language Natural Language Formal Language: are designed for computing purposes. Execution Python executes the code line by line Python tries to request memory from os Purpose of Datatypes Datatypes indirectly are used to allocate memory Programming Languages are of two types depending on how they deal with datatypes Static Typed Language: We need to be… Continue reading Gen-AI Developer Classroom notes 14/Sep/2025

Naive RAG Cons Multi hop & Multi queries might not work. Solving these problems also gathering live info from various datasources such as databases, api calls etc need Agentic RAG How do i verify if the RAG is working correctly. There are frameworks such as RAGAS, Truelens … Opensource Frameworks Enterprise Frameworks We need to… Continue reading Gen-AI Developer Classroom notes 13/Sep/2025

System Setup for python OS Package Managers Linux: Ubuntu: apt apt-get RedHat: dnf yum Mac: Homebrew Windows: winget Softwares: Git: Mac brew install git Windows: Download manually and install to get terminal option or use winget to configure terminal on your own python 3.13 Mac brew install python@3.13 Windows winget install –id=Python.Python.3.13 -e Visual Studio… Continue reading Gen-AI Developer Classroom notes 13/Sep/2025

Foundations contd Problem : Prime number Conversation Hi sara number = 9 index = 2 result = Prime until index < number if number % index == 0 result = Not Prime exit index = index + 1 say result Remember above steps as is_prime Usage Hi sara find is_prime 5 => Prime find is_prime… Continue reading Gen-AI Developer Classroom notes 11/Sep/2025

Foundations contd Project Euler Equals == Not Equals != Problem 1 Refer Here for problem 1 Conversation Hi sara Remember 9 as max Remember 1 as start Remember 0 as result do the follwing steps till start == max if start % 3 == 0 or start % 5 == 0 then add start to… Continue reading Gen-AI Developer Classroom notes 10/Sep/2025

Programming Foundations contd… Overview even number or not Convesation Hi sara Remember 70 as number calculate number % 2 and remember as result if result is zero say even else not even Approximate python code number = 70 result = number % 2 if result == 0: print("even") else: print("not even") Approximate Java code int… Continue reading Gen-AI Developer Classroom notes 09/Sep/2025

Python Programming for Generative AI Developer Goals: Solve Enterprise usecases for GenAI Integrations Design Patterns/Architectural Patterns Unit Testing Application types Thick Client Command line applications Desktop Application / Mobile Apps Thin Client Web Applications Programming It is about instructing the right way to get the things done.

Transformers To perform language translations & other nlp activities RNN’s were used which used to process text sequentially. Transformers were introduced with attention concept Refer Here for gpt2 visualization. Datasets used in LLAMA training CommonCrawl Wikipedia Books Github Code ArXiv Once the model is trained with all of the data, this model is referred as… Continue reading Gen-AI Developer Classroom notes 04/Sep/2025

ML Trainings Supervised Learning Prediction Unsupervised Learning Clustering NLP (Natural Language processing): This is the ability of system to understand natural language Terms: Input Layers Hidden Layers Output Layers Backpropgation Weights Biases Activation functions Embedding: Convert word to vector. Embedding Model: A model which has all possible words with vector values Understanding Text For models… Continue reading Gen-AI Developer Foundations Classroom notes 03/Sep/2025

Problems to solve Source (kbarticles) are prone to updates, which means that the vector database has to be updated How can i know whether my RAG is giving quality results. MultiModal Document Types Next : RAG Types Agentic RAG Graph RAG Handling the Source updates (files/folders) While indexing, we read files and lets find a… Continue reading Gen-AI Developer Classroom notes 02/Sep/2025

LLM (Large Language Models) Most of us would have used applications of LLMS such as ChatGPT Gemini Claude Perplexity At its very core LLM is a Neural Network which predicts the next token(Word) Aritificial Intelligence is a field of study to simulate human brain/behavior/actions etc.. AI was tried few times historically and it failed this… Continue reading Gen-AI Developer Classroom notes- Foundations – 02/Sep/2025

Cloud Services for Generative AI All the cloud providers offers services for enabling Generative AI Application Development. They offer services where the models are hosted and we can use models from cloud. Embeddings Vector Databases Cloud Storage integrations Deployments Services offered: Google: Vertex AWS: Bedrock Azure: Azure OpenAI, Azure AI Foundry Google Vertex Unified AI… Continue reading Gen-AI Developer Classroom notes 31/Aug/2025

Semantic Chunking When we have long, well-structured manuals, policies, text books, how-to knowledge bases, we need meaning preserving chunks for enhanced queries Langchain has an experimental feature called Semantic Chunker Refer Here

Chains in Langchain Runnable Sequence Refer Here for api docs Runnable Sequence can be simplified with pipe Maintaining Prompt file templates Lets explore load_prompt Refer Here for the usage of yaml files for storing templates Give semantic versioning

Practical RAG contd.. We have indexed the documents and stored in vector database Now lets create the generation pipeline Prompting techniques A Basic approach to create a well defined prompt is Give a persona Add your context Ask the question Add restrictions (optional) Add output format (optional) Prompt for home remidies As a grandmother from… Continue reading Gen-AI Developer Classroom notes 28/Aug/2025

Practical RAG Lets build a RAG for question and answering over KB Articles Refer Here for the data in raw form Install the following packages uv add langchain langchain-community langchain[google-vertexai] tqdm unstructured libmagic langchain-text-splitters langchain-chroma For loading the data we will be using directory loader Refer Here for the notebook with further steps.

RAG Generation Pipeline When a user asks a question, we retrieve the data from vector database and add it to the prompt, pass it across to LLM to get a response In the below image block in green color refers to indexing pipeline and blue refers to generation pipeline. The steps: user query will be… Continue reading Gen-AI Developer Classroom notes 25/Aug/2025