Master Agentic AI for QA + Build Custom Agents - APR'26

Discover Python as a high-level language, learn its uses from automation and QA testing to data analytics and AI, and master eight practical modules from basics to advanced topics.

Learn to install Python on mac using Homebrew, verify Python 3 and pip, create Python and pip aliases, and run a quick test to confirm a working setup.

Install Python on Windows by visiting python.org, selecting the stable Windows installer, and adding Python to the environment path; verify with python --version and pip --version.

Learn how to install PyCharm on Mac, choose between brew or DMG installation, and use PyCharm Community Edition for Python development.

Learn how to install PyCharm on Windows by downloading the official JetBrains installer for Windows, choosing the appropriate exe, and completing the standard setup.

Discover how agentic ai autonomously set goals, observe and analyze, plan with multi-step reasoning, generate test cases and automation scripts, and execute tests with memory and tool integration.

Learn how agentic AI addresses QA pain points like test authoring, regression triage, and log analysis with autonomous, 24/7 agents; explore real use cases and available tools.

Large language models are trained on trillions of text and code to generate language, code, and summaries for agentic AI, while handling tokens and context windows in QA.

Choose the right LLM for QA tasks by weighing paid and open-source options, context windows, and prompt engineering, while applying best practices to optimize cost and accuracy.

Master prompt engineering, the art of crafting clear, specific prompts to guide generative and agentic AI toward accurate outputs, by testing and refining with focus on clarity, context, and tone.

Learn about seven common prompt types—instructional, contextual, formatting, open-ended, specific example, clarification, and comparative—and how to choose the right type for accurate AI results.

This live demo shows how instructional, contextual, formatting, open-ended, and comparative prompts shape real-time AI responses using an LLM, with banking and e-commerce examples.

Explore prompt framework essentials, including zero-shot, few-shot, chain-of-thought prompting, plus reason plus act (react), and structured json outputs for qa agents.

Install Ollama on macOS via two methods: download the DMG from olama.com and drag it to the applications folder, or run brew install ollama from the terminal.

Start OLAMA on your Mac via Spotlight or Terminal, then verify it runs through the UI or localhost:11434, and manage it with start, stop, and --version commands.

Install OLAMA on Windows by downloading the official Windows exe from the OLAMA site, running the setup with default or custom locations, and launching the OLAMA UI.

Start OLAMA on your Windows machine from the start menu, then open the web UI and verify it runs at localhost:11434 after using OLAMA start in the command prompt.

Learn to download OLAMA models, pull or run them locally, and use OLAMA list to manage models and run prompts on your machine.

Choose the OLAMAS model for QA tasks by balancing reasoning versus coding, hardware constraints, and model size, favoring LLAMA 3.1/3.2 or QUEN 3 for reasoning and DeepSeek coder for coding.

Explore how model parameters, measured in billions, affect knowledge and reasoning, and how to choose between 1.3B, 6.7B, 33B, or enterprise models for local QA tasks (7–8B recommended).

Hit Ollama with curl via the API and /api/generate endpoints, passing the model and prompt to obtain a JSON response and learn to stream or parse with jq.

Learn to call Olama with curl using the slash api chat endpoint for multi-turn conversations, and learn roles such as user and system to structure prompts.

Explore the olama model library to pick a model for qa tasks, coding, or reasoning, noting no one best model and recommending llama 3.1/3.2 or quen 3 for reasoning.

Generate a named API key on the OpenAI API platform (project or org), copy and save it securely, and use it for authenticated requests.

Save the OpenAI API key on your local machine and set it in your path or environment. Start with a $5 credit and use cheap models like 4.1 nano.

Learn to call the OpenAI API with curl, hit the chat completion endpoint using a bearer API key, and extract the JSON content with jq or JSON path.

Learn the three roles in the OpenAI API—user, developer, and assistant—and how they set prompts, priority, and context across conversations.

Generate and save the Google Gemini API key, create a new project, and set the key in your zshrc or environment variable, mirroring the OpenAI process.

Learn to connect to the Gemini API via curl using the Gemini API key in the xgoog header, and extract text responses with jq.

Install and verify Python on macOS with Homebrew, selecting the latest stable version (3.14.2) and testing with a simple print hello. The lecture also covers Windows installation steps and checks.

Understand how Python virtual environments isolate per-project dependencies and manage different versions. Learn to create and activate a venv on Mac or Windows with python -m venv.

Install core dependencies efficiently using a requirements.txt file that pins exact versions for five packages—pandas, python.env, httpx, pydantic, and tenacity—and verify installation with pip list.

Install the OpenAI, Gemini, and Olama dependencies to enable LLM interactions in your Python project, using a requirements file and pip installation checks.

Perform a quick Python smoke test to verify dependencies like pandas by writing a CSV into an outputs folder with a small script.

Learn to set up a .gitignore in the project root to prevent .env and other sensitive files from being pushed to GitHub, with common rules for Python caches and IDEs.

Set up a standardized project folder structure for agent development, including data, outputs, and src with core and agents, plus env files and git-tracked folders.

Learn to set up a .env and .env.example to manage secrets and configurations, including provider, model, and keys like OpenAI and Google API, with easy provider switching.