Document Q&A RAG chatbot
Build a chatbot that answers questions from a custom document collection using FAISS vector store, LangChain retrieval chain, and streaming responses via FastAPI.
LangChainFAISSFastAPI
Home / Courses / Agentic AI & Gen AI
Agentic AI & Gen AI
From prompt engineering to LangGraph multi-agent systems, the complete course for building production Generative AI and Agentic AI applications with LangChain, RAG, and LangSmith.
Curriculum
Three months covering Gen AI foundations, LangChain and RAG, agentic workflows with LangGraph, and production deployment with LangSmith.
Wk 1: LLM fundamentals, transformer architecture (conceptual), tokens, context windows, temperature, top-p; OpenAI API, Anthropic API, calling models programmatically
Wk 2: Prompt engineering, zero-shot, few-shot, chain-of-thought, system prompts, structured output with Pydantic; function/tool calling
Wk 3: Local models with Ollama (Llama 3, Mistral, CodeLlama); comparing cloud vs local; embeddings, what they are and how to generate them; cosine similarity
Wk 4: LangChain fundamentals, chains, LCEL (LangChain Expression Language), prompts, output parsers, Runnable protocol
Wk 5: Vector stores, FAISS, ChromaDB, Pinecone; document loaders, text splitters, embedding pipelines
Wk 6: RAG end-to-end, retrieval chain, contextual compression, hybrid search (keyword + semantic), re-ranking
Wk 7: Conversation memory, BufferMemory, SummaryMemory, VectorStoreMemory; multi-turn conversational RAG; LangSmith tracing setup
Wk 8: AI agents, ReAct pattern, tool definitions, agent executors; custom tools (Python functions, APIs, web search)
Wk 9: LangGraph, nodes, edges, state management, conditional branching, cycles; building a research agent with LangGraph
Wk 10: Multi-agent workflows, supervisor agents, sub-agents, parallel execution; Human-in-the-Loop patterns; LangSmith evaluation and debugging
Wk 11: Deploying LangChain apps with FastAPI, streaming with Server-Sent Events, authentication, rate limiting, Docker packaging
Wk 12: LangSmith production monitoring, evaluation datasets, regression testing; capstone agentic application, mock interviews
Projects
Build a chatbot that answers questions from a custom document collection using FAISS vector store, LangChain retrieval chain, and streaming responses via FastAPI.
Build a ReAct agent that can search the web, read URLs, run Python code, and synthesise findings, using LangChain tools and an Anthropic or OpenAI LLM backend.
Design a multi-agent system using LangGraph where a supervisor routes tasks to specialist agents (researcher, writer, reviewer) and enforces human approval before final output.
A fully deployed, production-grade agentic application with a FastAPI backend, streaming UI, LangSmith monitoring, and a Docker container, ready to present to an employer.
Career Outcomes
Gen AI and Agentic AI engineers command some of the highest fresher salaries in Pune's current IT hiring cycle.
Four end-to-end AI applications you built and deployed, the most effective way to demonstrate Gen AI skills to employers.
A TechSummitTrainers completion certificate for this program once you finish the course and capstone project.
Roles this prepares you for
Course FAQ
Generative AI produces content from a prompt. Agentic AI goes further: an AI agent can plan, use tools, call APIs, search the web, and loop over its own output to reach a goal with minimal human input. Building agentic systems is the frontier of applied AI engineering.
Solid Python fundamentals, functions, classes, file I/O, working with libraries like requests and json. If you have done TechSummitTrainers' full-stack Python or Data Science course, or equivalent experience, you are ready. No deep ML background is required.
RAG (Retrieval-Augmented Generation) connects an LLM to your own documents and databases. It is now the dominant architecture for enterprise Gen AI applications, from internal chatbots to document intelligence systems.
LangChain is the most widely adopted open-source framework for building LLM applications. It provides abstractions for chaining calls, connecting to tools and APIs, and building retrieval pipelines, dramatically reducing the boilerplate needed for production AI applications.
LangGraph is built on LangChain for building stateful, multi-step AI agents using a graph-based execution model. It allows agents to loop, branch, and backtrack, enabling complex, reliable agentic workflows that simple chains cannot handle.
OpenAI (GPT-4o), Anthropic (Claude), Google (Gemini), and open-source models via Ollama (Llama 3, Mistral). The LangChain abstraction layer means the same code works across providers.
Gen AI Engineer, AI Application Developer, Prompt Engineer, LLM Engineer, AI Solutions Architect. These are the fastest-growing and highest-paying entry-level roles in India's current hiring cycle, with fresher salaries in Pune ranging from ₹5 LPA to ₹14 LPA.
The course fee is discussed during your free demo call. We offer flexible payment options. Contact us to know more.
Yes. This course focuses on building applications with LLMs and AI agents, not training models from scratch. Python programming and basic familiarity with APIs are the only real prerequisites.
LangSmith is the observability platform for LangChain applications. It lets you trace every LLM call, debug chains and agents, run evaluation datasets, and monitor production systems. It is taught as the standard tool for production AI quality assurance.
Yes, and this is very popular among working software developers who want to add Gen AI skills. Batch timings are confirmed at your free demo.
Any degree with solid Python programming experience, BE/B.Tech (CS/IT), BCA, BSc IT, or equivalent practical Python skills from industry or other courses.
Agentic AI refers to AI systems that can autonomously plan, reason, use tools, and execute multi-step tasks to achieve a goal, with minimal human direction at each step. While a standard Generative AI application answers a prompt and stops, an AI agent loops: it thinks about what to do next, calls a tool (search the web, run code, query a database, call an API), observes the result, and continues until the task is complete. This shift from "AI as assistant" to "AI as autonomous worker" is transforming how software companies build products, and creating urgent demand for engineers who can build these systems.
In Pune's hiring market, the demand for Gen AI and Agentic AI skills has grown faster than any other technical specialisation in 2024–25. Product companies, IT services firms, and startups are all actively hiring engineers who can integrate LLMs into applications, build RAG systems, and design agentic workflows.
These three terms describe distinct levels of an AI application stack. Large Language Models (LLMs) are the foundation, neural networks trained on massive text datasets that can generate human-quality text, code, and structured output in response to a prompt. Generative AI is the broader category of AI systems that produce new content, text, images, audio, video, including but not limited to LLMs. Agentic AI is an application architecture on top of LLMs: an agent uses an LLM as its reasoning engine, but wraps it in a planning loop that can call external tools, store memory, and take actions in the world. This course covers all three layers, from the LLM API layer up through LangChain, RAG, and LangGraph agentic workflows.
LangChain is the most widely adopted open-source framework for building production applications with large language models. Without LangChain, every LLM application requires significant boilerplate: managing prompts, handling streaming, connecting to vector databases, implementing retry logic, managing conversation history, and chaining multiple calls together. LangChain provides clean abstractions for all of these, including LCEL (LangChain Expression Language), which lets you compose complex pipelines in a few lines of code. It has become the de-facto standard in Indian companies adopting Gen AI.
RAG (Retrieval-Augmented Generation) solves the core limitation of LLMs: their training data has a cutoff date, and they know nothing about your specific business context. With RAG, you embed your company's documents into a vector database (FAISS, ChromaDB, Pinecone), retrieve the most relevant passages when a user asks a question, and inject them into the LLM's context window alongside the question. The result: an AI assistant that can accurately answer questions about your internal documents, your product knowledge base, your support tickets, or any other domain-specific information.
Simple LangChain chains work well for linear pipelines, input → process → output. But real agentic systems need to branch, retry, loop, and coordinate multiple agents. LangGraph solves this with a graph-based execution model: you define nodes (actions or LLM calls) and edges (transitions, including conditional branches). An agent can loop back when it needs more information, spawn sub-agents for parallel work, or pause to wait for human approval. LangGraph is the tool used in production agentic systems at scale, and it is the core skill that separates an AI application developer from a prompt engineer.
TechSummitTrainers is located at Office No. 509, Kushal Wallstreet, Fergusson College Road, Shivajinagar, central Pune. This course is taught by practitioners who have built and deployed LangChain and LangGraph applications on live projects. Every lab session involves writing and running real code against actual LLM APIs, not simulated environments. The hands-on project structure means you graduate with four deployed AI applications in your portfolio: the most effective way to demonstrate Gen AI skills to a hiring manager in Pune's current market.
Book your free demo today. No payment, no commitment. Next batch starts Nov 10, 2026.
Get a Call Back →