AIEngineeringTraining

Start with Python installation, development environments, variables, data structures, conditionals, loops, functions, libraries and basic terminal use.

Work on Streamlit installation, web-based interface development, core methods and application publishing.

Get to know JSON and XML structures. Integrate your Python applications with external services and API sources.

Work on database creation, table structures, CRUD operations and basic querying logic.

Learn the concepts of classes, objects, methods, constructors, inheritance and encapsulation.

Gain data analysis practice from vector and matrix operations to DataFrame structures, and from CSV, Excel and SQL data to data preprocessing.

Create different chart types with Matplotlib and turn analysis results into visual output.

Learn the Django architecture, the model-template-view structure, the admin panel and dynamic application development.

Work on Scikit-learn, regression models, Logistic Regression, Decision Tree, Random Forest, KNN and prediction models.

Take your programming and data analysis foundation to a more advanced level before deep learning.

Learn the basic structure of artificial neural networks, weights, the concept of bias and how they work.

Examine the Sigmoid, Softmax, ReLU, Tanh, Linear and Leaky ReLU functions through hands-on examples.

Build deep learning models with TensorFlow and Keras. Apply the concepts of batch, epoch, optimizer and regularization.

Examine the CNN architecture through image processing problems and gain model development practice.

Build models by learning how RNN and LSTM architectures work for sequential data.

Take the models you build into application interfaces that can interact with users.

The AI agent concept, what is an LLM? What is an embedding? What is a token? Differences between a chatbot and an AI agent. What is the RAG architecture? What are tool calling and function calling? The difference between a workflow and an agent. Backend architecture in LLM applications. Differences between cloud models and local models. API-based model usage. Balancing speed, cost and accuracy when choosing a model. The security approach in agent applications. The general architecture of corporate AI applications.

Using the OpenAI API. OpenAI chat models. The OpenAI response structure. OpenAI embedding models. OpenAI function calling logic. Using OpenAI tool calling. Using the Gemini API. Gemini LLM models. Gemini embedding models. Using Gemini function calling. Getting structured output with Gemini. Open-source model integrations. Qwen and similar open LLM models. How running a local model works. Using models with Ollama. Serving models with LM Studio. API key management. Using environment variables. Model selection and usage scenarios. Using more than one model in the same project.

The embedding concept. Text embedding logic. Choosing an embedding model. OpenAI embedding models. Gemini embedding models. Open-source embedding models. Multilingual embedding usage. Embedding quality with Turkish data. Converting text into vectors. Cosine similarity logic. Calculating similarity scores. Chunking strategies. Text splitting methods. Extracting text from PDF files. Preparing data from TXT, HTML and JSON sources. Using metadata. Document indexing logic. Updating and deleting embedding data.

What is a vector database? Which problems are vector databases used for? Installing ChromaDB. The ChromaDB collection structure. Storing embeddings with ChromaDB. Semantic search with ChromaDB. Using metadata in ChromaDB. The ChromaDB local development setup. Installing Qdrant. Running Qdrant with Docker. Qdrant collection logic. The Qdrant point and payload structure. Storing embeddings with Qdrant. Semantic search with Qdrant. Filtering in Qdrant. Qdrant scoring logic. Comparing Qdrant and ChromaDB. Using a vector database in production. Backup and migration logic.

What is RAG? The Retrieval-Augmented Generation architecture. When is a RAG system used? The document loading flow. Data cleaning. Creating chunks. Generating embeddings. Writing to the vector database. Generating an embedding from the user query. Retrieving similar documents. Building the context. Asking the LLM a question with context. How answers with sources are produced. Techniques for reducing hallucination. Prompt design. The multi-document RAG structure. Building a PDF-based question-answering system. A web page-based RAG system. Building a corporate document search assistant. Methods for improving RAG performance.

What is LangChain? LangChain use cases. Using PromptTemplate. Model wrapper structures. Using output parsers. Getting structured output. Chain logic. Using retrievers. Building a RAG pipeline. Defining tools. The memory concept. Using conversation memory. OpenAI integration with LangChain. Gemini integration with LangChain. Using open models with LangChain. Error handling. Retry mechanisms. Folder structure in LangChain projects.

What is LangGraph? Why use LangGraph? Agent workflow logic. The node structure. The edge structure. State management. Using conditional edges. Building multi-step agents. Designing agents that use tools. Memory-backed agent flows. The human-in-the-loop approach. Supervisor agent logic. Building a RAG agent with LangGraph. Building a decision tree with LangGraph. Error handling inside an agent. Fallback flows for wrong answers. Agent debugging logic. Multi-agent structures. Using LangGraph in real business scenarios.

What is function calling? What is tool calling? Creating a tool schema. JSON Schema logic. Introducing tools to the LLM. Using Python functions as tools. Writing a tool that queries a database. Writing a tool that calls an external API. Writing a tool that reads files. Writing a tool that performs calculations. Integration logic with CRM, ERP and accounting systems. Authorization and security checks. Restricting tool use by user permission. Tool result validation. Preventing incorrect tool use. Letting the agent act on external systems.

FastAPI project structure. Creating API endpoints. Building a chat endpoint. Building a RAG endpoint. Building an agent endpoint. Request and response models. Data validation with Pydantic. Streaming response logic. API key checks. A simple auth structure. CORS settings. Environment settings. LangChain integration with FastAPI. LangGraph integration with FastAPI. Connecting FastAPI to Qdrant. Connecting FastAPI to ChromaDB. Preparing the service with Docker. Production deployment logic.

Prompt engineering fundamentals. System prompt logic. The distinction between user prompts and developer prompts. Steering agent behavior. Preparing structured prompts. What is prompt injection? Taking precautions against prompt injection attacks. Data leakage risks. API key security. User-based authorization. Testing LLM answers. Evaluating agent outputs. Basic evaluation logic. Logging and error tracking. Tracking token usage. Cost control. Measuring model answer quality.