Developing Generative AI Applications on Azure

Learn Python basics as a high-level interpreted, open-source language renowned for readability, with libraries for web development, data analysis, and AI, plus setup across Windows, macOS, and Linux.

Explore Python data types including integers, floats, strings, and booleans, and learn type conversion and operators, including arithmetic, comparison, logical, bitwise, membership, and identity, for hands-on labs.

Explore Python lists and tuples through examples, covering indexing, slicing, and operations like append, extend, pop, remove, reverse, and list comprehensions, then compare mutable lists with immutable tuples and unpacking.

Explore dictionaries and key-value pairs, including access, updates, and iteration with items. Practice file handling, built-in and user-defined functions, and variables, using Colab notebooks for hands-on Python practice.

Explore Python operators through hands-on demonstrations in Colab, defining variables, performing arithmetic operations, and printing results with comments for clear documentation.

Demonstrates Python comparison operators—equal, not equal, greater than, less than, greater than or equal to, and less than or equal to—through print statements and hands-on coding.

Explore Python operators with a hands-on demo that evaluates x, y, and z comparisons using print statements, including greater than, less than or equal to, and not equal.

Explore Python data types in an interactive demo, covering integers, floats, strings, and complex numbers, with hands-on exercises in Colab or Jupyter and printing types to reinforce usage.

Explore Python strings, including single-line and multi-line literals, printing, newline escapes, and slice notation; practice with variables name age marks and printing their data types.

Review assignment one in Python data types by printing name, score, and age, and display their data types as integer, float, and string.

Demonstrate Python lists with a product catalog, using zero-based and negative indexing and performing append, delete, and insert operations to manage items like laptop, smartphone, and tablet.

Learn practical Python list operations in a hands-on demo, including extend, append, and slicing, with real-world examples like new arrivals, an inventory printout, and adding accessories.

Revisit a Python list assignment demo showing append to add a charger and extend to add screen protector and phone case, then print the updated product list.

Explore Python tuples versus lists, highlighting mutability, memory efficiency, and use cases for static inventories, then experiment with tuple operations and access patterns in a hands-on coding demo.

Explore Python tuples, including insertion via the plus operator, extending with new items, and slicing in a shopping inventory, while contrasting tuples with lists for read-only, static data.

Learn to use Python dictionaries to model product inventory with key-value pairs, update stock, compute total inventory value, and identify low-stock items.

Learn to manage inventory with Python dictionaries by removing zero stock items, merging inventories with quantity sums for overlapping products, and filtering by category such as apparel, footwear, and accessories.

Review Python dictionary operations by filtering footwear items from an inventory, print product and quantity, then update jeans and sneakers quantities, and practice in a notebook.

Learn Python file handling from creating and writing to text files, reading and appending, to building and reading csv files with csv writer and csv reader.

Download data from URLs, fetch with http requests, read CSV with csv.reader and StringIO, and build a log processor to extract errors for machine learning workflows.

Explore python functions, from inbuilt tools like length, max, and min, to defining your own function and using input to create dynamic greetings. Create reusable code for apps.

Learn to define Python functions for calculating total price from price and quantity, handle float and integer inputs, display product details, and apply tax on totals.

Explore how to declare and update Python variables, print their values, and inspect data types such as integers, floats, and strings using type().

Learn string concatenation with variables to join first name and last name, perform basic arithmetic with Python operators, and compute subtotal, discount, and final total in an e-commerce store.

Learn how to define item price and quantity in Python, compute a subtotal, apply a 10% discount, and print subtotal, discount amount, and final total.

Explore the fundamentals of machine learning, including supervised, unsupervised, and reinforcement learning, models and neural networks, data preparation, model evaluation, and bias and privacy.

Explore how machine learning powers voice assistants, autonomous vehicles, and personalized recommendations, while highlighting Generative AI, efficiency at scale, pattern recognition, and continuous improvement.

Learn supervised learning with labeled data and model training, and compare it to unsupervised and reinforcement learning, including regression, image classification, and testing data.

Master unsupervised learning with unlabeled data that discovers hidden patterns through clustering and association, using examples like customer segmentation and promotional offers, with TensorFlow applications on Azure.

Explore how reinforcement learning trains an agent to navigate an environment through exploration. Identify rewards, negative feedback, and policy as key mechanisms behind safer, efficient autonomous vehicle decisions.

Identify the differences between supervised, unsupervised, and reinforcement learning using labeled versus unlabeled data and agent actions. See examples like spam classification, sentiment analysis, image classification, and autonomous vehicles.

Explore neural networks inspired by the human brain, powering deep learning, and see how artificial intelligence and machine learning use input, hidden, and output layers to recognize patterns.

Discover how deep learning powers real-world applications like speech recognition, image recognition, and natural language processing, from Android voice search to Facebook tagging and ChatGPT text generation.

Explore how decision tree algorithms perform classification and regression with an intuitive, human-like decision process, powering bank risk analysis, e-commerce customer segmentation, and medical diagnosis applications.

Explore linear regression, a supervised learning method for predictions from labeled data, using simple and multiple variable models to forecast prices and premiums with tools like scikit-learn, pandas, and numpy.

CNNs are deep neural networks that analyze visual imagery and learn features for translation-invariant recognition. They enable image and video recognition, facial recognition, and medical imaging using TensorFlow, Keras, MXNet.

Explore recurrent neural networks (RNNs) and their memory-based sequential data handling for real-time language translation, sentiment analysis, and speech recognition, using TensorFlow, Keras, PyTorch, and MXNet.

Explore four ML algorithms—linear regression, decision tree, CNN, and RNN—and their use cases in stock price prediction, loan approval, customer segmentation, image recognition, translation, and time-series data.

Improve model predictions by prioritizing data quality through rigorous data preparation, cleaning, and normalization. Learn techniques for handling missing values, feature encoding, and train-test splits to boost Azure AI applications.

Evaluate machine learning models by understanding overfitting and underfitting, balancing training and testing accuracy, and using cross-validation to stop training at the right fit.

Explore ethics in machine learning by mitigating bias, promoting fairness, protecting privacy and security, and ensuring transparency and accountability in model predictions.

Explore the future of machine learning as AI integrates into everyday devices, Copilot in office apps, and autonomous vehicles. Address ethical AI, explainability, and societal impacts, including job transformation.

Explore numpy basics by creating arrays from lists, initializing zeros and ones, using arange and reshape, practicing arithmetic, indexing, slicing, and computing min, max, and sum.

Master pandas basics for machine learning data handling: create data frames from lists, define columns, save and load CSV files, select columns, and add a discount price column.

Learn to create a NumPy array and convert it to a pandas dataframe. Save and read csv files to demonstrate the connection between NumPy and pandas for machine learning data.

Learn to build a supervised image classifier with TensorFlow and Keras using a labeled cats and dogs dataset, training and validation splits, and a convolutional neural network.

Train a supervised image classifier by rescaling and resizing training and validation data using flow from directory to 150 by 150 images, with batch size 20 across 15 epochs.

Save the trained model with model.save to a .h5 file, then load it with TensorFlow Keras load_model; preprocess a test image to 150 by 150 and add a batch dimension.

Load the trained TensorFlow model, upload an image, and predict cat or dog with a CNN. Resize to 150 by 150, normalize 0-1, and improve accuracy over 15 epochs.

Generate synthetic e-commerce data in Colab using NumPy and pandas to illustrate unsupervised learning, including customer names, purchases, categories, cities, browsing times, and views.

Explore an e-commerce data set built with pandas and numpy, creating data frame of customers and purchases, then apply k-means clustering on spend levels and summarize category and city sales.

Master linear regression with a single variable to predict insurance premiums from age, using colab, numpy, and sklearn; fit data and predict ages 21 and 48.

Demonstrates multi-variable linear regression to predict insurance premiums from age, height, and weight, with data cleaning through mean imputation, model fitting, and predicting outcomes.

Explore what generative AI is, its applications across sectors on Azure, and the behind-the-scenes tools and libraries, with ethical, privacy, legal, security, and project management considerations.

Explains generative AI as a class of AI that creates content from data, tracing origins from 1950s–60s foundations to learning and neural networks enabling text, image, video, and music generation.

Explore a hands-on generative AI demo that generates text, images, music, or video from a single prompt, using ChatGPT, Leonardo AI, and Soundraw.io.

Explore how generative AI powers content creation, personalized education, healthcare, gaming, fashion, advertising, finance, law, robotics, urban planning, retail, and customer service.

Demonstrates generating a complete video from a single prompt using a free AI video tool, with background music, voiceover, fonts, and scenes tailored for YouTube and Facebook ads.

Explore neural networks (CNNs and RNNs), GANs, VAEs, transformer models, diffusion and autoregressive models, reinforcement learning, and fine tuning to power generative AI on Azure.

Examine the ethics, privacy, and legal aspects of generative AI, covering fairness, accountability, transparency, and data protection laws like GDPR and CCPA, with bias and privacy case studies.

Learn to generate complete PowerPoint slides from a single prompt using gamma, customize the outline, themes, images, and animations, and export as PDF or PowerPoint.

Define scope and objectives of genai projects, identify data needs, set measurable milestones, monitor progress with dashboards and agile methods, and lead cross-functional teams toward value in healthcare and retail.

Explore real-world gen ai security risks by examining adversarial, evasion, poisoning, privacy, and backdoor attacks across input and training data, and learn the importance of clean data and robust safeguards.

Explore the future of AI, focusing on automation, job displacement, and upskilling for adaptive workers with gen AI literacy, data analysis, and evolving machine learning capabilities.

Generate AI voices for studio-quality voiceovers in a single prompt, across nearly 100 languages with male and female options, customize voices, and download the audio.

Explore how generative AI supports fraud detection, predictive modeling, cash flow forecasting, portfolio optimization, automated reporting, and intelligent process automation in finance, with multilingual chatbots and personalized offers.

Leverage GenAI for personalized product recommendations, predictive lead scoring, and data-driven marketing campaigns that optimize timing and conversions in ecommerce.

Explore how AI in HR management enables applicant screening, resume parsing, and attrition prediction, delivering personalized assessments, automated performance feedback, and data-driven, unbiased evaluations.

Demonstrates generating ai avatar videos by cloning yourself, selecting an avatar, scripting with chatgpt, and delivering ai news anchor style broadcasts for a cloud enabled channel.

Discover how generative artificial intelligence enables personalized medicine and early diagnosis, enhances hospital efficiency, and powers computer vision for medical image and video recognition in robotic surgery and bronchoscopy.

Master prompt engineering to craft effective prompts, refine them iteratively, and leverage data-driven techniques to unlock AI capabilities and improve output quality.

Master zero-shot prompting, few-shot prompting, and chain-of-thought prompting to improve outputs in Azure generative AI, using context-based, iterative prompts and prompt optimization.

See how a submitted prompt activates a language model, encodes input into tokens, and initializes the model, then token generation uses probability estimation and vocabulary lookup to produce output.

Explore prompt engineering with ChatGPT-3.5 through hands-on demos on generating content, adding constraints, translating prompts, and using prompts to act as different roles for building scripts, plans, and learning exercises.

Compare ChatGPT-3.5 and ChatGPT-4, showcasing web search, attachments, and rapid prompts. Demonstrate practical prompts to summarize PDFs, generate code, create CSV data, and craft social media content.

Explore prompting with the Anthropic cloud, compare its up-to-date data to GPT-3.5, and learn practical prompts for Python, code generation, optimization, debugging, and building web apps.

Explore prompting with google gemini to write code, translate text, and generate a blog on AI in transportation within the google ecosystem.

Explore the fundamentals of natural language processing, including natural language understanding and generation, key applications, evolution, challenges, and the pipeline from data to model evaluation.

Explore practical NLP applications, from language detection and translation to smart replies, voice assistants, spell check, customer service bots, and spam filtering, with a hands-on spam-filtering demo in Python.

Trace the evolution of natural language processing from a heuristic approach to machine learning and deep learning with neural networks, highlighting keyword search, rule-based parsing, regular expressions, and expert systems.

Learn machine learning methods for NLP, from supervised and unsupervised learning to deep learning with transformers, including Naive Bayes, SVMs, decision trees, random forest, and sentiment analysis.

Explore key NLP challenges such as ambiguity, contextual word meaning, irony and tone, spelling errors, and language diversity, including dialects and slang, that affect interpretation and processing.

Explore nlp tasks, including text classification and sentiment analysis, using ml models like naive bayes, svm, random forest, and transformers; cover information retrieval and natural language queries with real-world applications.

Detects input language with character frequency and n-grams to route text through the right pipeline, then covers statistical and neural translation along with natural language understanding and natural language generation.

Explore text summarization and topic modeling, including effective and abstractive summaries and LDA, NMF, and HDP, plus text generation and speech to text applications across media, academia, and business.

Explore the six stages of an NLP pipeline, from text acquisition and pre-processing to feature extraction, application integration, and model evaluation, with data import, web scraping, and APIs.

Explore an end-to-end NLP pipeline for sentiment analysis of Amazon iPhone reviews, from web text acquisition and preprocessing to feature extraction and classification with BERT.

Explore text classification for email filtering with a Python workflow that preprocesses emails, builds features via a count vectorizer, trains a spam detector, and tests samples.

Demonstrate a Colab workflow to install nltk and sumy, upload a notebook, and summarize a large text into three concise lines using an LSA summarizer.

Learn NLP data preprocessing by cleansing text from diverse sources, removing emojis, URLs, HTML tags, and punctuation. Lowercase text, correct spelling, and filter stopwords to refine data for downstream analysis.

Explore the fundamentals of large language models, their applications, advantages and limitations, plus ethical considerations. Learn to fine-tune, customize, and prototype multi modal LMs with Hugging Face for programmatic access.

Explore how large language models train on vast unlabeled data using transformers, and leverage pre-training and fine-tuning with examples like GPT three, BERT, and T5.

Discover how large language models enable image and video generation, music creation, text generation, translations, sentiment analysis, code generation, and chatbots.

Master custom and fine tuning language models by defining objectives, data collection and pre-processing, curating Java data from GitHub, and training or fine-tuning on targeted tasks.

Explore how multimodal LLMs process and integrate text, images, and audio to generate a single end-to-end solution, from scripts and visuals to audio, for education, healthcare, and advertising.

Access OpenAI programmatically using Python in Colab to build generative AI applications, then submit a prompt to the GPT 3.5 turbo chat completion endpoint and view the response.

Explore OpenAI NLP tasks in a hands-on demo, including text summarization, translating to Spanish, text completion, prompts with temperature, and building a simple chatbot using Python and GPT-3.5.

learn to access the anthropic cloud programmatically with python in Colab, configure api keys and endpoints, and test prompts using the client.message_create method to build AI applications.

Demonstrates NLP tasks on entropy cloud by programmatically summarizing text, translating to French, and building a cloud-enabled chatbot using anthropic APIs in Python with an Azure AI Studio workflow.

Fine-tune distilbert base uncased on telecom message data to classify harmful, spam, and generic messages using Hugging Face transformers and trainer, with Colab training and a saved fine-tuned model.

Develop an OpenAI chatbot with a Flask web interface using Python and an OpenAI key, guided through Colab and a standard templates and static directory.

Upload a PDF to the chatbot, then ask questions to summarize or extract information from the document using a Flask app powered by Python, OpenAI, and PDF Plumber.

Build a ChatGPT bot that accepts text prompts, queries OpenAI for responses, converts them to speech with a text-to-speech tool, and plays the audio on the user's device.

Demonstrate prerequisites for Native World chatbot by provisioning a VM (AWS, Azure, or Google Cloud) with two CPUs, four RAM, 40 disk, and installing Python 3 and pip.

Deploy a web chatbot on an EC2 Linux VM by uploading the notebook via Colab and running a Docker-based Streamlit app with a Postgres vector database.

This demo guides you to set up an OpenAI API key, crawl a target website, index content with a vector DB, and validate the chatbot's knowledge.

Demonstrates building a Telegram chatbot powered by OpenAI, including creating a Telegram bot, obtaining API keys, installing dependencies, and a Python app using GPT-3.5 for responses.

Explore Hugging Face and its transformers library, accessing pre-trained NLP models like GPT-2 for text generation, text classification, sentiment analysis, and named entity recognition through pipelines and datasets.

Explore how transformers and LM models enable NLP tasks like text generation, sentiment analysis, translation, and summarization using Hugging Face pipelines.