Generative AI Developer
A 14-week hands-on course that takes learners from Python basics to working generative AI applications. Learners build with LLM APIs from OpenAI, Anthropic and Google, Hugging Face models, embeddings and vector databases, RAG document Q&A, LangChain, LlamaIndex, LangGraph, function calling and agents, FastAPI and Streamlit apps, fine-tuning basics, evaluation and guardrails, and deployment with Docker and cloud. Every module ends with a small project for a Surat client. The course builds portfolio-ready skill; it does not guarantee a job or income.
- Duration: 14 weeks
- Modules: 14
- Fee: Rs 39,999
- Field: AI & Emerging Technology
Roles this track prepares you for
- Generative AI Developer
- Junior AI Engineer
- LLM Application Developer
- AI Automation Developer
- RAG Developer
- AI Chatbot Developer
- Python Developer
- Prompt Engineer
- AI Engineer
Curriculum
- Python Foundations for AI Developers - Learners set up Python and learn variables, decisions, loops, functions, modules, lists, dictionaries and strings, the base for every later AI project.
- Setting Up Python and Running Your First Program
- Variables, Data Types and Operators
- Decisions and Loops
- Functions and Modules
- Lists, Dictionaries and Strings
- Python for Real Projects: Files, APIs and Git - Learners read and write files, handle errors, use virtual environments, call REST APIs and work with Git and GitHub like a developer.
- Reading and Writing Files, JSON and CSV
- Errors, Exceptions and Debugging
- Virtual Environments and Packages
- HTTP, REST APIs and the Requests Library
- Git and GitHub Workflow
- LLM Fundamentals and Prompt Engineering - Learners understand tokens, context windows, settings and hallucinations, then write, structure, test and version prompts that behave reliably.
- How LLMs Work: Tokens and Context Windows
- Model Settings, Limits and Hallucinations
- Prompt Patterns: Roles, Instructions and Examples
- Structured Outputs and JSON Prompts
- Testing and Versioning Prompts
- Working with LLM APIs: OpenAI, Claude and Gemini - Learners call LLM APIs from Python, keep keys safe, manage conversations, stream and structure outputs, handle errors and costs and compare providers.
- API Keys, Security and Your First Call
- Chat Messages, System Prompts and Conversations
- Streaming and Structured Output
- Errors, Rate Limits and Cost Control
- Comparing Providers and Switching Models
- Hugging Face and Open Models - Learners find and choose models on the Hugging Face Hub, run pipelines, decide between open models and hosted APIs and check licences, bias and safe use.
- The Hugging Face Hub: Finding and Choosing Models
- Transformers Pipelines: Text, Sentiment and Summaries
- Open Models or Hosted APIs: How to Choose
- Model Licences, Bias and Safe Use
- Running Models Locally and in Free Cloud Notebooks
- Embeddings and Vector Databases - Learners understand embeddings and similarity search, use FAISS, Chroma and Pinecone, and prepare data with chunking, metadata and filters.
- What Embeddings Are and Why They Work
- Similarity Search: Cosine and Nearest Neighbours
- FAISS and Chroma Hands-On
- Pinecone and Managed Vector Databases
- Chunking, Metadata and Filters
- RAG Architecture and Document Q&A - Learners build the full retrieval-augmented generation pipeline: loading and chunking documents, retrieval, prompting with citations, better retrieval and debugging failures.
- The RAG Pipeline from Document to Answer
- Loading and Chunking Documents
- Retrieval, Prompts and Citations
- Improving Retrieval: Reranking and Hybrid Search
- Why RAG Fails and How to Debug It
- LangChain and LlamaIndex Frameworks - Learners use LangChain and LlamaIndex to build chains, memory, parsers and query engines, and learn when a framework helps and when plain code is better.
- Why Use a Framework and When Not To
- LangChain: Prompts, Models and Chains
- LangChain: Memory, Output Parsers and Tools
- LlamaIndex: Indexes and Query Engines
- Framework or Plain Code: Making the Call
- Function Calling and AI Agents - Learners let models call tools, design tool schemas, build the agent loop and add permissions, human approval and limits so agents stay safe.
- Function Calling: Letting a Model Use Tools
- Designing Tools and Schemas
- The Agent Loop: Think, Act, Observe
- Agent Safety: Permissions and Human Approval
- Agent Failure Modes and Limits
- LangGraph and Multi-Step Agent Workflows - Learners model workflows as graphs with state, routing, checkpoints and human approval, and learn when multi-agent designs help and when they add risk.
- Graphs, State, Nodes and Edges
- Routing and Conditional Edges
- Memory, Checkpoints and Persistence
- Human-in-the-Loop Workflows
- Multi-Agent Patterns and When to Avoid Them
- FastAPI and Streamlit for AI Apps - Learners wrap AI features in FastAPI endpoints with validation, add keys and rate limits, build Streamlit interfaces and test and document their APIs.
- FastAPI Basics: Endpoints and Validation
- Serving an LLM Feature as an API
- Streamlit: Building a Simple Interface
- API Keys, Authentication and Rate Limiting
- Testing and Documenting Your API
- PyTorch Basics and Fine-Tuning Basics - Learners learn tensors and a training loop, decide between prompting, RAG and fine-tuning, prepare data and understand LoRA and parameter-efficient tuning.
- Tensors and Autograd Basics
- A Training Loop for a Small Model
- Prompting, RAG or Fine-Tuning: Which One?
- Preparing Data for Fine-Tuning
- LoRA and Parameter-Efficient Tuning Concepts
- AI Evaluation, Guardrails and Safety - Learners build evaluation sets, use metrics and LLM judges with care, defend against prompt injection, protect personal data and monitor AI apps in use.
- Building an Evaluation Set
- Metrics and LLM-as-Judge, and Their Limits
- Prompt Injection and Defences
- Privacy, PII and Moderation Guardrails
- Logging, Monitoring and Feedback Loops
- Docker, Cloud Deployment and Capstone - Learners containerise AI apps with Docker, manage secrets, deploy with AWS or Azure basics, watch cost and latency, then build and present a capstone AI application with a GitHub portfolio.
- Docker Basics for AI Apps
- Secrets, Config and Environments
- Deploying to AWS or Azure: The Main Steps
- Cost, Latency and Monitoring in Production
- Capstone: Build and Deploy an AI Application
- GitHub Portfolio and Interview Preparation for AI Roles
Frequently asked questions
How long is the Generative AI Developer course?
It runs for 14 weeks, with 14 modules. It is self-paced online learning from EasySkill Career Academy, Surat.
How much does the Generative AI Developer course cost?
The fee is Rs 39,999.
What will I build in Generative AI Developer?
Every module ends in a graded practical task that AI checks in about 30 seconds, so you finish with a portfolio you can show to employers or clients.
Which roles does Generative AI Developer prepare me for?
It prepares you to apply for roles such as Generative AI Developer, Junior AI Engineer, LLM Application Developer, AI Automation Developer, RAG Developer. It does not guarantee a job or income.
Can I learn online, or do I have to visit a centre?
You can learn fully online. EasySkill also has centres in Adajan, Vesu and Bhestan, Surat, if you prefer to visit.
All career tracks