Course Foundation

#

Course Duration and Scope:
This course is designed to be delivered over 27 sessions, each lasting approximately 1 to 1.5 hours. It offers a comprehensive, step-by-step journey from the fundamentals of AI chatbots to advanced machine learning and deep learning techniques. The curriculum spans five levels covering practical use of popular AI chatbots and APIs, local AI model deployment with secure networking, sophisticated Arabic NLP combined with web scraping, multimodal image processing and emotion recognition, and foundational deep learning architectures and algorithms. Each session integrates theory with practical examples and assignments, culminating in projects that synthesize learned skills for real-world AI applications.

Related pages to execute this track

#

1. Python Data Engineering & MLOps roadmap for full phase-by-phase progression.
2. Courses hub for complementary tracks.
3. Workshops & Camps for practical live sessions.
4. Mentorship Services for one-to-one implementation and project strategy.

License and Disclaimer

#

This course material is designed and provided solely for use within Neurobotics Academy.
All rights are reserved by Neurobotics Academy and the course developer Eng. Mulham Fetna, CEO & Founder of Neurobotics Academy.

Course Design and Maintenance:
This course has been designed, developed, and maintained by Eng. Mulham Fetna and Neurobotics Academy as part of their STEM curriculum development initiatives.

Date of Development:
December 2025

Rights Reserved:
All intellectual property rights, including copyrights, are exclusively reserved for Neurobotics Academy and Eng. Mulham Fetna. Unauthorized use, reproduction, or distribution outside of Neurobotics Academy is strictly prohibited.

Weeks 1-2: Getting Started with AI Chatbots and APIs

#

• Introduction to AI chatbots, their types, and applications
• Hands-on demos with ChatGPT, Google Gemini, Claude, Copilot
• Understanding API concepts and authentication
• Building simple chatbot apps using paid and free APIs (OpenAI, Grok, Hugging Face)
• Assignment: Build and compare chatbots with free and paid APIs

Weeks 3-4: Local AI Models and Secure Networking

#

• Installing and running local AI models (LLaMA, Ollama, OpenWebUI)
• Exposing local models via REST APIs with FastAPI
• Networking basics: localhost, IP addresses, ports
• Secure remote access using tunneling tools (Twingate, Cloudflare Tunnel)
• Security best practices: zero-trust models, API keys, rate limiting

Weeks 5-6: Arabic Natural Language Processing (NLP) and Web Scraping

#

• Arabic NLP foundations: AraBERT, CAMeL Tools, PyArabic
• Tokenization, Named Entity Recognition (NER), and sentiment analysis for Arabic text
• Web scraping tools and techniques (BeautifulSoup, Selenium) for Arabic content collection
• Integrating scraped data into Arabic NLP pipelines
• Assignment: Build an Arabic social media sentiment analyzer

Weeks 7-8: Image Processing and Multimodal Emotion Detection

#

• Image manipulation basics with OpenCV: resizing, filtering, color spaces
• Object detection using pretrained YOLO models on images and videos
• Facial emotion recognition using DeepFace and FER models
• Combining text and image sentiment for multimodal analysis
• Assignment: Develop multimodal social media post analysis system

Weeks 9-12: Deep Learning Frameworks and Advanced Machine Learning

#

• Neural network fundamentals: perceptrons, activations, forward and backward passes
• Loss functions and backpropagation explained with examples
• Supervised learning: regression, classification, decision trees
• Unsupervised learning: clustering (K-Means, DBSCAN), dimensionality reduction (PCA, t-SNE)
• Deep architectures: CNNs, RNNs, LSTMs, transfer learning
• Comparative analysis of algorithms: strengths, weaknesses, and best use cases
• Final project review and integration across domains

Additional Recommendations

#

• Hands-on Practice: Encourage building projects incrementally to reinforce concepts
• Community Engagement: Join AI and ML communities for peer support and collaboration
• Version Control: Introduction to Git and GitHub for code management and sharing
• Ethics and Responsible AI: Discuss implications and safe AI deployment practices

Summary Table

#

This course is crafted and delivered by Eng. Mulham Fetna, CEO & Founder of Neurobotics Academy, dedicated to fostering cutting-edge AI & ML expertise for students and professionals.

Course Objectives

#

Course Structure Overview

#

Your “AI/ML from Zero to Hero” course spans five levels, progressing from accessible cloud AI to advanced custom models, with 20-30 sessions of 1-1.5 hours each. Allocate roughly 4-6 sessions per level to fit the range, ensuring in-depth coverage: partial explanations with code snippets, integrated real-world examples (e.g., building a customer support bot), and pre-session assignments with hints. End each level with a review session, project assignment (e.g., deploy a multi-API chatbot for Level 1), and culminate in a graduation project like an end-to-end Arabic sentiment analysis app with web scraping and image emotion detection.

Level 1: Famous AI Chatbots and APIs

#

Start with user-facing demos of ChatGPT (versatile for writing/coding, GPT-4o model with voice/tools), Gemini (Google-integrated, multimodal with real-time search/video), Copilot (productivity in Microsoft tools), Claude (long-context reasoning), and Meta AI (casual social use). Key differences include ChatGPT’s broad features versus Gemini’s deep research and Claude’s structured outputs. Transition to APIs: Compare paid OpenAI/Gemini APIs for apps like translators, then free options like Grok API (xAI’s witty, real-time responses) and Hugging Face (open models like Llama via inference API).

Level 2: Local Models and Networking

#

Cover running Llama models locally via Ollama or LLM Studio for privacy/offline use, with OpenWebUI for a ChatGPT-like interface. Dive into local APIs (e.g., Ollama endpoints), networking basics (expose via localhost tunneling), and tools like Twingate (secure zero-trust access) or Cloudflare Tunnel (free, easy port forwarding) to access your local AI across devices.[8]

Levels 3-5: Advanced Topics

#

Level 3 (NLP): Use Arabert/CAMeL Tools/PyArabic for Arabic text analysis (tokenization, NER, sentiment); integrate web scraping with BeautifulSoup/Selenium for real-time data (e.g., analyze social media posts).

Level 4 (Images): Pretrained YOLO for object detection, OpenCV for processing, plus sentiment/emotion models (e.g., FER via DeepFace) on faces/text overlays.

Level 5 (Frameworks): TensorFlow/PyTorch basics—neural nets (CNNs/RNNs), loss functions (MSE/cross-entropy), regression/classification, supervised (labeled data, high accuracy but data-intensive) vs. unsupervised (clustering, scalable but less precise)—with pros/cons tables per algorithm.

Session and Progression Tips

#

Each session: 20-min partial concepts (e.g., API auth code), 30-min integrated example (e.g., RAG chatbot with LangChain), 20-min assignment (build/deploy with hints like “use streamlit for UI”). Reviews consolidate via quizzes/projects; track via GitHub for continuity with your bulk-processing workflow. This fits 25 sessions (5/level), scalable to 30 with extras on RAG/agents.

Course Plan

#

Level 1: Famous AI Chatbots and APIs (5 Sessions)

#

Session 1: Introduction to AI Chatbots

#

• Overview of AI chatbots: What and why?
• Demo popular AI chatbots as users (ChatGPT, Google Gemini, Claude, Copilot)
• Key differences in design, use cases, features, and architecture concepts
• Assignment: Compare chatbot responses on a given topic with reasoning hints

Session 2: Using Chatbot APIs (OpenAI, Google, Anthropic)

#

• Introduction to API concepts and authentication
• Making API calls to ChatGPT and similar platforms
• Simple chatbot app example with API calls (Python + simple UI)
• Assignment: Build a simple Q&A using ChatGPT API with hints

Session 3: Exploring Free AI APIs (Grok, Open-source APIs)

#

• Overview of free AI API platforms (Grok, Hugging Face, Cohere)
• Demo API calls and capabilities vs paid APIs
• Integrate a free API into a mini chatbot or text generation app
• Assignment: Create a chatbot using a free API and document limitations

Session 4: Integrating Multiple APIs and Advanced Features

#

• Combining APIs for enhanced features (e.g., RAG - retrieval augmented generation)
• Handling errors, rate limits, and streaming results
• Use-case example: Info bot with multiple API sources
• Assignment: Extend chatbot to query multiple APIs with hints

Session 5: Level 1 Review and Project Assignment

#

• Recap key concepts and compare chatbot types and APIs
• Level project assignment: Build a customer support chatbot combining paid and free APIs
• Guidance on project milestones and evaluation criteria

Level 2: Local Models and Networking (5 Sessions)

#

Session 6: Introduction to Local AI Models (LLaMA, Ollama, OpenWebUI)

#

• Understanding local models vs cloud models
• Installing and running LLaMA-based models offline
• Demo OpenWebUI for chatting locally
• Assignment: Set up a local model instance with basic interaction

Session 7: Working with Local Model APIs

#

• Exposing local models as REST APIs
• Simple backend server creation for local AI model calls
• Assignment: Build a small local API wrapper for a LLaMA model

Session 8: Networking Basics for Local AI Access

#

• Intro to networking concepts: localhost, ports, IPs
• Using tunneling tools: Twingate, Cloudflare Tunnel
• Secure access to local AI services from other devices
• Assignment: Connect to a local AI model remotely using tunneling

Session 9: Security and Best Practices for Local AI Deployment

#

• Security principles: zero trust, encryption, firewall
• Practical security setup using Twingate, Cloudflare
• Assignment: Harden local AI API deployment and test secure access

Session 10: Level 2 Review and Project Assignment

#

• Review local model setup, APIs, networking
• Project: Deploy a local chatbot accessible securely from multiple devices
• Guidelines and expectations

Level 3: NLP for Arabic and Web Scraping (5 Sessions)

#

Session 11: Arabic NLP Foundations (AraBERT, PyArabic, CAMeL Tools)

#

• Arabic script challenges and preprocessing
• Tokenization, POS tagging, NER with AraBERT and CAMeL
• Assignment: Perform named entity recognition on sample Arabic text

Session 12: Sentiment and Emotion Analysis in Arabic

#

• Sentiment analysis models and techniques
• Applying pre-trained sentiment models to Arabic datasets
• Assignment: Build a basic Arabic sentiment classifier

Session 13: Web Scraping for Data Collection

#

• Tools overview: BeautifulSoup, Selenium
• Scraping Arabic web data for NLP preprocessing
• Assignment: Write a scraper for Arabic news articles

Session 14: Integrating NLP Pipeline with Scraped Data

#

• Pipeline overview: scraping → cleaning → analysis
• End-to-end example with Arabic text analysis on scraped data
• Assignment: Complete a mini pipeline project with hints

Session 15: Level 3 Review and Project Assignment

#

• Review Arabic NLP, scraping basics
• Project: Analyze sentiments and entities in Arabic social media posts
• Deliverables and guidelines

Level 4: Image Processing and Emotion Detection (5 Sessions)

#

Session 16: Image Processing Basics with OpenCV

#

• Image tasks: resizing, filtering, color spaces
• Hands-on OpenCV scripts
• Assignment: Build an image preprocessor

Session 17: Object Detection with YOLO

#

• YOLO architecture and inference
• Running YOLO models on custom images/videos
• Assignment: Implement object detection and generate reports

Session 18: Emotion and Sentiment Detection from Images

#

• Emotion datasets and models (FER, DeepFace)
• Sentiment analysis from facial expressions
• Assignment: Build a simple emotion detector on sample images

Session 19: Integrating Image Processing and NLP Results

#

• Multimodal analysis combining text and images
• Example: Analyze social media posts with text and image sentiment
• Assignment: Build a multimodal mini-project

Session 20: Level 4 Review and Project Assignment

#

• Recap image processing, detection, sentiment analysis
• Project: Social media post analyzer combining image and text emotion detection
• Assessment criteria

Level 5: Neural Networks and Advanced ML (5-7 Sessions)

#

Session 21: Neural Network Fundamentals

#

• Perceptrons, activation functions, forward pass
• Simple neural net example in TensorFlow/PyTorch
• Assignment: Build a perceptron for binary classification

Session 22: Loss Functions and Backpropagation

#

• MSE, cross-entropy explained
• Backpropagation illustrated with examples
• Assignment: Implement custom loss function

Session 23: Supervised Learning Techniques

#

• Linear/logistic regression, decision trees overview
• Training and evaluation practice
• Assignment: Build a supervised classifier from scratch

Session 24: Unsupervised Learning Techniques

#

• Clustering, dimensionality reduction
• Visualization with PCA, t-SNE
• Assignment: Cluster an unlabeled dataset

Session 25: Deep Learning Architectures

#

• CNNs, RNNs, LSTMs overview
• Transfer learning with pretrained models
• Assignment: Fine-tune a CNN on a sample dataset

Session 26: Pros and Cons of Algorithms, Use Cases

#

• Comparative analysis of algorithms
• When to use supervised vs unsupervised, deep vs shallow
• Assignment: Choose algorithms for problem statements and justify

Session 27: Final Course Review and Graduation Project Introduction

#

• Review core concepts from all levels
• Present graduation project guidelines: e.g., Arabic text and image sentiment app with deployment
• Q&A and next steps