Business-Ready AI with Google Vertex AI (Easy and No Code)

Set up a Google Cloud Console account from scratch on cloud.google.com to access Vertex AI, including creating an organization profile, verifying payment details, and starting the free trial.

Set up Vertex AI with no-code tools, enable required APIs, and configure Vertex AI Studio to build your first app, manage billing, and explore the dashboard.

Explore vertex ai studio's multimedia tools, including chat, images, video, music, and live, and generate eight-second videos from a Dragon Ball prompt with seed and safety options.

Design your first AI app on Vertex AI without coding by setting the system message, adding personality, tweaking behavior, and testing live to build a custom AI assistant.

Deploy your first app by building an email writer, configuring prompts and variables, and deploying a no-code web app on Google Vertex AI with Cloud Run.

uncover how the system message defines who the AI is and how it talks, shaping responses through structure, rules, and roles.

Embed the system message in the app, compare prompts with and without instructions, and use few shot prompting with high quality data in a no-code Vertex AI workflow to evaluate results.

Explore zero-shot, one-shot, and few-shot prompting alongside persona and goal prompting for large language models, showing how patterns and roles boost accuracy and consistency.

Explore how to update a no-code AI app using few-shot examples, defining input and output, and deploying prompts in a web app to craft engaging emails.

Explore jailbreaking risks in no-code Vertex AI apps, revealing how system messages, few-shot examples, and data leakage can occur, and learn defensive measures for enterprise AI security.

Explore how to hide and test the system prompt across six levels of jailbreaking in a no-code ai chat app, using a Harry Potter character to reveal secrets.

Keep the momentum by filling out the next lecture's feedback form, sharing what's firing you up and what's missing the mark to turbocharge the next half of the course.

Explore no-code retrieval augmented generation with Vertex AI, using your own documents to ingest, index, retrieve, and generate trustworthy answers that pull exact excerpts through governance and vector databases.

Rag, or retrieval augmented generation, grounds answers with up-to-date data by retrieving relevant documents, augmenting prompts, and generating grounded responses across healthcare, finance, education, and more.

Create a corpus in Vertex AI using the rack engine with the basic tier, upload PDFs, and configure chunking, embeddings, and default parsing for multilingual data.

Chunking splits text into smaller pieces so the AI searches the most relevant chunks, using small chunks for fact-based questions and larger chunks for broader context and reasoning.

Embeddings turn words into numbers so computers can search, classify, and power chatbots and document search, with related meanings ending up near each other in context.

Explore how vector stores serve as an AI memory, turning words, images, and sounds into embeddings for meaning-based search and relevant results.

Enable the Google Drive API and grant viewer access, then import data in Vertex AI from Google Drive or local files, with practical debugging steps for import errors.

Test and validate your Rag Engine corpus in Vertex AI Studio, inspect the restaurant data, and deploy the app to overcome quota limits and resource exhaustion.

Debugging RAG in Google Vertex AI without code by configuring storage roles, choosing supported regions for rack engine, and testing prompts to generate recommendations and deploy mini web apps.

Improve your rag by applying system instructions and role constraints for concise, friendly answers, using documents when unsure. Leverage lovable to build a restaurant bot site with free credits.

Diogo highlights how data solves complex business challenges, from sales planning and a/b tests to his startup Joint Beta that analyzes data to optimize menus and pricing.

Deploy a digital waiter using Vertex AI and Lovable to power a no-code front door, handling gluten-free pasta questions, reservations, and wine pairings with real-time responses.

Craft a complete prompt to design a restaurant website with an embedded bot, using Vertex AI, lovable, and no-code workflow, guided by menu materials and screenshots.

Enable cloud for the project to activate backend, set up a bottom-right chatbot with Gemini API key and Google Cloud service account, and address Vertex AI rag engine considerations.

Debug the chatbot, diagnose edge function errors, resolve Google Cloud service account authentication, and manage credits while planning a one-go prompt to finish the integration.

Finalize the chatbot for a no-code workflow by tracing deployment logs, access tokens, and production vs development gaps, then improve color contrast and accessibility through iterative prompts.

Learn to build machine learning models with Vertex AI and no-code AutoML, turning your data into predictions from a dashboard, training a model, and using it in apps.

Explore data sets and training in Google Vertex AI, focusing on tabular data for a smartphone price multi-classification problem, uploading CSV files, creating a dataset, and generating statistics.

Train a new model using Vertex AI AutoML tabular training for price range classification, exploring feature selection, correlation with target, and setting training budgets and early stopping.

AutoML acts as a data science copilot, testing models and tuning automatically to pick the best one based on performance for your data.

Evaluate classification models using accuracy, confusion matrix, precision, recall, and F1, with examples of true and false positives and negatives, and consider ROC AUC for imbalanced datasets.

Use precision, recall, F1, and AUC ROC over accuracy for imbalanced data, guided by the confusion matrix and business costs of false positives and false negatives.

Learn batch inference on Google Vertex AI with an end-to-end workflow: train AutoML tabular models, configure batch predictions from a CSV dataset, and monitor results with email alerts.

Explore no-code model evaluation with batch inference, using model registry to inspect the precision-recall area under the curve of 0.99, f1 scores, and threshold tradeoffs.

Audit Vertex AI predictions with no-code AutoML by exporting results, previewing CSVs, and visualizing predictions. Handle batch inference status and exit-handler errors to ensure reliable, business-ready insights.

Celebrate finishing the course and showcase business-ready AI with Google Vertex AI in a no-code workflow. Share your certificate on LinkedIn and post projects to build a portfolio.