🚀 Generative AI Bootcamp (12 Hours | 2 Days) –
Focused on Business Applications
👥 Audience: Business Students (Non-technical, but with basic Python understanding)
🛠 Mode: Hands-on, using Google Colab, Hugging Face, and Streamlit
🗓️ Duration: 2 Days (6 hours per day)
📌 Target Business Domain: Financial Services & Market
Analysis
✅
Project: AI-Powered Financial Market Assistant
This assistant helps financial analysts and business managers by summarizing reports,
✅
analyzing market trends, and answering queries using Generative AI.
✅
Users can upload financial reports (PDFs) and query insights directly.
It integrates RAG (Retrieval-Augmented Generation) to fetch relevant data before
✅
answering.
The final product is a web-based AI assistant (Streamlit UI) deployed on Hugging Face
Spaces.
🗓️ Day 1: Foundations & Prompt Engineering (6 Hours)
🔹 Session 1: Introduction to Generative AI & LLMs (1 Hour)
● What is Generative AI? How does it impact businesses?
● History of LLMs: Evolution from GPT-2 → GPT-4, Claude, Gemini, Mistral
● How LLMs process text: Tokens, embeddings, attention
● Business applications: Chatbots, market forecasting, personalized finance
🛠 Hands-on:
✅ Run a simple LLM-based chatbot using OpenAI API in Google Colab
🔹 Session 2: Exploring LLMs for Business Use (1 Hour)
● Overview of Key LLMs: OpenAI (GPT-4), Anthropic (Claude), Google Gemini, Mistral,
LLaMA
● Proprietary vs Open-source LLMs: When to use what?
● Cost vs Latency vs Accuracy trade-offs
● How companies like BloombergGPT are leveraging LLMs for finance
🛠 Hands-on:
✅ Load and compare open-source LLMs from Hugging Face using transformers
🔹 Session 3: Prompt Engineering Basics (1.5 Hours)
● Prompt Engineering Fundamentals
○ Zero-shot, Few-shot, Chain of Thought (CoT)
○ Role of Temperature & Max Tokens in responses
○ Business use cases: Writing financial reports, generating summaries
🛠 Hands-on:
✅ Compare zero-shot vs few-shot prompting
✅ Generate executive summaries from financial data
🔹 Session 4: Retrieval-Augmented Generation (RAG) for Business Reports
(1.5 Hours)
● Why do LLMs hallucinate?
● How RAG fixes hallucination with real-time data retrieval
● Vector databases: FAISS, ChromaDB, Pinecone
✅
● Business use cases:
✅
Extracting insights from financial reports
Answering queries based on company filings
🛠 Hands-on:
✅ Implement RAG using FAISS & OpenAI API
✅ Query business documents using a custom RAG pipeline
🔹 Session 5: Agentic AI & LLM Agents for Market Analysis (1 Hour)
● What is an AI agent?
● How LLM-based agents work (LangChain, AutoGPT, BabyAGI)
✅
● Business use cases:
✅
Autonomous financial research
Automated trading bots
🛠 Hands-on:
✅ Use LangChain to build an AI agent that generates investment insights
🗓️ Day 2: Building & Deploying an AI-Powered Financial
Market Assistant (6 Hours)
🔹 Session 6: End-to-End Project – AI Market Analysis Assistant (3 Hours)
✅ Problem: Business analysts spend hours reading company filings & reports to extract key
✅ Solution: Build an AI assistant that:
financial insights.
● Accepts PDF uploads (financial reports, earnings calls)
● Uses RAG (FAISS) to extract insights
● Generates AI-powered summaries
● Answers business-related queries
🛠 Hands-on Steps:
1️⃣ Frontend UI (Streamlit)
● Create a simple Streamlit UI for uploading PDFs
● Display chat history and responses
2️⃣ Backend (RAG Implementation)
● Use FAISS to index financial reports
● Implement retrieval pipeline using OpenAI / Mistral
3️⃣ LLM-Based Query Handling
● Generate business insights from reports
● Answer investor questions with relevant financial data
✅ Final Output: A working AI Financial Market Assistant
🔹 Session 7: CI/CD & Automating Deployment (2 Hours)
● Introduction to CI/CD for LLM apps
● Pushing code from Google Colab to Hugging Face Repo
● Automating deployment using GitHub Actions
🛠 Hands-on:
✅ Colab Notebook → Hugging Face Repo (Push updates automatically)
✅ Test and refine the deployed assistant
🔹 Session 8: Final Wrap-up & Q&A (1 Hour)
● Challenges in AI adoption for business
● Best practices for maintaining an AI-powered assistant
● Future trends in Generative AI & Finance
🛠 Tools Used
✔ Google Colab (hands-on coding)
✔ Hugging Face (LLMs & deployment)
✔ OpenAI API / Mistral / LLaMA (financial analysis)
✔ FAISS / ChromaDB (for RAG)
✔ Streamlit (for UI)
✔ GitHub CI/CD (for automation)
🔥 Key Takeaways
✅ Hands-on experience with LLMs, RAG, and AI Agents
✅ Build and deploy a real-world financial assistant
✅ Automate deployment with CI/CD to Hugging Face
🚀
This bootcamp ensures that business students not only learn about Generative AI but also
build and deploy an AI-powered Financial Market Analysis Assistant.
