I love this chart on AI skills. I regret that I was reading this on my phone and didn’t catch the source to cite it (you know, fumble fingers), so I apologize in advance. If you have the citation, please leave it in the comments. While I explore how I got to the final chart at the bottom, you can always skip to the AI Learning Roadmap at the end.

AI Skills and How To Achieve Them

One of the challenges for me in considering AI is how to identify skills beyond simple prompt engineering or designing prompts. People say, “You’ve got to learn AI,” but then after that, there’s not a lot of detail. Even when we talk about things like standards-based tools like the TCEA Essential Learning Expectations, those focus on educational applications in the classroom around topics such as:

• AI Literacy
• Ethical AI Use
• AI-Enhanced Learning
• AI and Critical Thinking
• AI Collaboration and Innovation

For example, in regards to AI and Critical Thinking, the TCEA ELEs speak to:

• Teach students to critically analyze AI-generated content.
• Guide students in verifying AI outputs using multiple sources.
• Develop AI-related problem-solving skills through classroom activities.

And that’s wonderful. But how do *I* develop the skills that are worth paying for in the marketplace? That’s where this chart comes in.

AI Skills

With that in mind, I asked ChatGPT to give me some practical suggestions on how to achieve specific AI skills. Those AI skills include:

• GenAI – use AI to generate text, images, and more.
• Artificial neural networks – build computer systems that learn like human brains do.
• Computer vision – teach computers to “see” and understand images.
• PyTorch (machine learning library) – use this tool to build powerful AI applications.
• Machine learning – teach computers to learn from data.
• Applied machine learning – use machine learning to solve real problems.
• Deep learning – build advanced AI systems for complex tasks.
• Supervised learning – train AI using labeled examples.
• Reinforcement learning – train AI through trial and error.
• Machine learning operations (MLOps) – manage and deploy machine learning systems effectively.

This didn’t make a lot of sense to me, so I asked Claude AI to make me a Mermaid Syntax flowchart.

Flowchart

As you can see, the flowchart below (generated with Mermaid.Live, aren’t they awesome?) moves from yellow (beginner) to blue (intermediate) to green (advanced). Expert and Deployment levels are displayed as red and black, respectively. I suspect that many new users of GenAI fall into yellow, but there’s a huge gulf between that and machine learning basics.

How This Flowchart Works

• Beginner (Yellow): Start with GenAI and Machine Learning Basics.
• Intermediate (Blue): Move to Supervised Learning and Applied Machine Learning.
• Advanced (Green): Learn Deep Learning, Neural Networks, Computer Vision, and Reinforcement Learning.
• Expert (Red): Work with specialized AI tools (PyTorch, OpenCV, OpenAI Gym).
• Deployment (Black): Deploy AI models using MLOps.

A Simple Table…Not the Final Draft

This is the table I started with, but it falls short of the flow chart since it doesn’t show the levels.

AI Skill & How to Achieve It	How You Can Achieve This
GenAI – use AI to generate text, images, and more. Example: Use tools like ChatGPT, DALL·E, or Midjourney to create stories, essays, or artwork.	Experiment with GenAI platforms, create AI-generated content for blog posts or presentations, and test AI-powered creative tools.
Artificial neural networks – build computer systems that learn like human brains do. Example: Use TensorFlow or PyTorch to build simple neural networks for tasks like handwriting recognition.	Take an online course on neural networks, experiment with no-code AI platforms like RunwayML, and test pre-trained AI models.
Computer vision – teach computers to “see” and understand images. Example: Use OpenCV to develop an image classification system, such as identifying objects in photos.	Try Google’s Teachable Machine for basic image recognition, follow OpenCV tutorials, and experiment with AI-based object detection apps.
PyTorch (machine learning library) – use this tool to build powerful AI applications. Example: Implement deep learning models for tasks like language translation or image generation.	Work through beginner-friendly PyTorch tutorials, join Kaggle competitions, and run small AI projects using cloud-based notebooks like Google Colab.
Machine learning – teach computers to learn from data. Example: Train a machine learning model in Python using Scikit-learn to predict housing prices.	Learn Python-based ML through platforms like Coursera or Udacity, work on small datasets, and use AutoML tools to build models without deep coding.
Applied machine learning – use machine learning to solve real problems. Example: Build a sentiment analysis tool to classify social media posts as positive or negative.	Find a problem you care about (e.g., analyzing student engagement), collect data, and use AI tools like Weka or Google AutoML to create solutions.
Deep learning – build advanced AI systems for complex tasks. Example: Develop a deep learning model for speech-to-text transcription using LSTMs or transformers.	Read books like Deep Learning with Python, test transformer models like GPT or BERT, and experiment with AI-powered apps like Whisper for speech recognition.
Supervised learning – train AI using labeled examples. Example: Train an AI model to recognize handwritten digits using the MNIST dataset.	Download datasets from Kaggle, use Scikit-learn to train models, and participate in AI challenges where you label and train data.
Reinforcement learning – train AI through trial and error. Example: Program an AI agent to play a simple game like Pong using reinforcement learning techniques.	Follow beginner reinforcement learning courses, use OpenAI Gym to test AI in game environments, and analyze AI decision-making processes.
Machine learning operations (MLOps) – manage and deploy machine learning systems effectively. Example: Deploy an AI model using cloud services like AWS or Google Vertex AI.	Learn the basics of cloud deployment with free-tier AWS/GCP accounts, use Streamlit to deploy AI apps, and practice with GitHub Actions for model updates.

The Final Table

Could the flow chart and the table below actually work?

Beginner Level

AI Skill & How to Achieve It	How You Can Achieve This
GenAI – use AI to generate text, images, and more. Example: Use tools like ChatGPT, DALL·E, or Midjourney to create stories, essays, or artwork.	Experiment with GenAI platforms, create AI-generated content for blog posts or presentations, and test AI-powered creative tools.
Machine learning – teach computers to learn from data. Example: Train a machine learning model in Python using Scikit-learn to predict housing prices.	Learn Python-based ML through platforms like Coursera or Udacity, work on small datasets, and use AutoML tools to build models without deep coding.

Intermediate Level

AI Skill & How to Achieve It	How You Can Achieve This
Applied machine learning – use machine learning to solve real problems. Example: Build a sentiment analysis tool to classify social media posts as positive or negative.	Find a problem you care about (e.g., analyzing student engagement), collect data, and use AI tools like Weka or Google AutoML to create solutions.
Supervised learning – train AI using labeled examples. Example: Train an AI model to recognize handwritten digits using the MNIST dataset.	Download datasets from Kaggle, use Scikit-learn to train models, and participate in AI challenges where you label and train data.

Advanced Level

AI Skill & How to Achieve It	How You Can Achieve This
Deep learning – build advanced AI systems for complex tasks. Example: Develop a deep learning model for speech-to-text transcription using LSTMs or transformers.	Read books like Deep Learning with Python, test transformer models like GPT or BERT, and experiment with AI-powered apps like Whisper for speech recognition.
Artificial neural networks – build computer systems that learn like human brains do. Example: Use TensorFlow or PyTorch to build simple neural networks for tasks like handwriting recognition.	Take an online course on neural networks, experiment with no-code AI platforms like RunwayML, and test pre-trained AI models.
Computer vision – teach computers to “see” and understand images. Example: Use OpenCV to develop an image classification system, such as identifying objects in photos.	Try Google’s Teachable Machine for basic image recognition, follow OpenCV tutorials, and experiment with AI-based object detection apps.
Reinforcement learning – train AI through trial and error. Example: Program an AI agent to play a simple game like Pong using reinforcement learning techniques.	Follow beginner reinforcement learning courses, use OpenAI Gym to test AI in game environments, and analyze AI decision-making processes.

Expert Level

AI Skill & How to Achieve It	How You Can Achieve This
PyTorch (machine learning library) – use this tool to build powerful AI applications. Example: Implement deep learning models for tasks like language translation or image generation.	Work through beginner-friendly PyTorch tutorials, join Kaggle competitions, and run small AI projects using cloud-based notebooks like Google Colab.

Deployment Level

AI Skill & How to Achieve It	How You Can Achieve This
Machine learning operations (MLOps) – manage and deploy machine learning systems effectively. Example: Deploy an AI model using cloud services like AWS or Google Vertex AI.	Learn the basics of cloud deployment with free-tier AWS/GCP accounts, use Streamlit to deploy AI apps, and practice with GitHub Actions for model updates.

There really seem to be two strands here, but what do I know? At this point, I’m still a white belt with AI, lacking experience in Python-based Machine Learning (and not likely to get it at this point). But I feel quite comfortable with GenAI.

AI Skill / Pathway	✅=No Code | ❌=Code Required
GenAI (Text & Image Generation) Using AI tools like ChatGPT, DALL·E, or Midjourney	✅
Basic Machine Learning Concepts Understanding how AI models work conceptually	✅
AutoML Tools (Google AutoML, Teachable Machine, RunwayML) Training AI models with minimal coding	✅
Supervised Learning with AutoML or No-Code AI platforms	✅
Applied Machine Learning Solving real-world problems using AI tools	✅
Computer Vision (Using pre-built image recognition tools like Teachable Machine)	✅
Deep Learning Using no-code platforms like Lobe.ai, Peltarion, or RunwayML	✅
Reinforcement Learning Using AI-driven simulations or pre-built AI models	❌
Building Custom AI Models Developing new architectures from scratch	❌
Using PyTorch, TensorFlow, or Scikit-learn for Custom AI Development	❌
MLOps Deploying AI models using no-code cloud tools like Google Vertex AI or AWS SageMaker Autopilot	✅
Full AI Development End-to-End AI Model Creation & Optimization	❌

With that chart of no code vs code, I have a good idea of how to proceed. What do you think?