Build a broad foundation
Understand major AI concepts before choosing a specialization such as ML, NLP, or vision.
Data & AI · Intelligent systems
Artificial Intelligence
Explore the foundations of Artificial Intelligence, including intelligent agents, search, reasoning, machine learning, natural language processing, computer vision, neural networks, generative AI, and responsible system design.
Build a practical understanding of how AI systems are framed, trained, evaluated, deployed, and reviewed across different application areas.
Course overview
Artificial Intelligence is a broad field focused on building systems that can perform tasks involving perception, prediction, language, reasoning, decision-making, and generation.
This course begins with the major ideas behind AI and then connects them to practical areas such as search algorithms, knowledge representation, machine learning, natural language processing, computer vision, and neural networks.
Later modules explore model evaluation, data quality, deployment considerations, generative AI concepts, and responsible AI practices.
AI is not one single technique. A good solution starts by defining the task, understanding the data, choosing an appropriate approach, measuring results, and reviewing risks before deployment.
Prerequisites
The course is designed to begin at an introductory level. The following knowledge will help you move through the practical activities more comfortably.
- Basic computer and file-management knowledge.
- Comfort with reading simple tables and charts.
- Basic mathematics such as percentages and averages.
- Willingness to work with structured data.
- Basic Python knowledge is recommended for coding exercises.
- Prior machine-learning knowledge is not required.
Readiness activity
Choose a simple prediction or classification problem, identify the input data and desired output, and write three possible ways the system could make a mistake. This introduces problem framing and risk awareness.
Who can explore this course?
Apply coding to intelligent systems
Connect Python, data preparation, model workflows, evaluation, and simple AI applications.
Understand model-based decisions
Explore features, training data, validation, metrics, and the limits of predictions.
Compare AI directions
Explore different AI domains and identify the subjects you want to study more deeply.
What you will learn
- Explain the main goals and subfields of AI.
- Describe intelligent agents and problem environments.
- Use search and optimization ideas to solve simple problems.
- Understand rules, logic, knowledge, and reasoning.
- Prepare data for basic machine-learning workflows.
- Compare supervised, unsupervised, and reinforcement learning.
- Explore NLP tasks such as classification and text processing.
- Understand computer-vision tasks and image features.
- Describe neural-network and deep-learning fundamentals.
- Evaluate models using appropriate metrics and test data.
- Identify risks involving bias, privacy, safety, and misuse.
- Build and document a practical AI application project.
Curriculum outline
The ten-module outline introduces AI broadly and then moves toward practical modeling, application development, evaluation, deployment, and responsible use.
01
Artificial Intelligence fundamentals
Build a shared vocabulary and understand the different ways AI systems can be designed.
- Definitions and historical directions of AI.
- Symbolic, statistical, and learning-based approaches.
- Narrow AI compared with general-intelligence concepts.
- AI systems, users, data, environments, and objectives.
- Examples across business, science, education, and services.
- Strengths, limitations, and common misconceptions.
Practice: Analyze three AI-enabled products, identify the task each performs, and describe what data or feedback the system may rely on.
02
Intelligent agents and problem solving
Explore how agents perceive an environment, choose actions, and work toward defined goals.
- Agents, environments, actions, and observations.
- Goals, utility, constraints, and decision quality.
- Fully observable and partially observable environments.
- Deterministic and uncertain outcomes.
- Problem formulation and state representation.
- Search spaces and practical solution limits.
Practice: Model a route-planning or scheduling problem with states, actions, costs, constraints, and a measurable goal.
03
Search, optimization, and reasoning
Study methods for exploring possible solutions and selecting useful actions under constraints.
- Breadth-first and depth-first search concepts.
- Uniform-cost and heuristic search ideas.
- Greedy search and A* intuition.
- Optimization and local-search concepts.
- Rules, facts, constraints, and logical inference.
- Trade-offs between speed, memory, and solution quality.
Practice: Implement or simulate a grid-navigation problem and compare two search strategies using path length and explored states.
04
Data preparation and machine learning workflow
Understand how raw data becomes a structured input for training and evaluating a model.
- Features, labels, examples, and target variables.
- Training, validation, and test data.
- Missing values and inconsistent records.
- Numerical scaling and categorical encoding.
- Data leakage and inconsistent preprocessing.
- Reproducible transformations and pipelines.
- Data quality and representative samples.
Practice: Prepare a small dataset, document its fields, identify missing values, and create a repeatable preprocessing workflow.
05
Supervised and unsupervised learning
Explore common learning settings and understand how a model learns patterns from data.
- Classification and regression problems.
- Clustering and grouping without labels.
- Nearest neighbors and tree-based models.
- Linear models and baseline comparisons.
- Overfitting, underfitting, and generalization.
- Feature selection and model complexity.
- Interpreting predictions carefully.
Practice: Train two baseline models on a small dataset, compare their validation results, and explain why one model may not generalize.
06
Natural language processing
Explore how systems process text and support tasks involving language, meaning, and communication.
- Text representation and tokenization concepts.
- Cleaning, normalization, and feature extraction.
- Text classification and sentiment analysis.
- Named entities and information extraction.
- Similarity, embeddings, and semantic relationships.
- Language ambiguity and context limitations.
- Evaluation and human review of language outputs.
Practice: Build a small text-classification experiment and review incorrect predictions for patterns in the data.
07
Computer vision and neural networks
Understand the core ideas behind systems that analyze images and learn layered representations.
- Pixels, channels, dimensions, and image labels.
- Classification, detection, and segmentation concepts.
- Feature extraction and visual patterns.
- Neurons, weights, activation functions, and layers.
- Training, loss, optimization, and validation.
- Convolutional-network intuition.
- Data augmentation and model limitations.
Practice: Design an image-classification project plan that identifies labels, data requirements, evaluation metrics, and possible errors.
08
Generative AI and modern AI applications
Explore systems that generate text, images, code, or other content and examine how applications are built around such models.
- Generative models compared with predictive models.
- Prompt structure, context, and output constraints.
- Retrieval and grounding concepts.
- Embeddings and semantic search.
- Tool use and workflow orchestration.
- Hallucination, verification, and human review.
- Privacy, copyright, and sensitive-data considerations.
Practice: Design a question-answering assistant that uses a small approved document set and includes source review and refusal behavior.
09
Evaluation, deployment, and monitoring
Learn how to measure model behavior and consider what happens after a system is placed into use.
- Accuracy, precision, recall, F1, and confusion matrices.
- Regression metrics and error analysis.
- Cross-validation and model comparison.
- Thresholds, calibration, and business costs.
- Inference workflows and API integration.
- Data drift, model drift, and monitoring concepts.
- Reproducibility, versioning, and rollback planning.
Practice: Create an evaluation report that includes metrics, examples of errors, limitations, and a recommendation for further data collection.
10
Responsible AI and capstone delivery
Review the wider impact of AI systems and prepare a project that documents goals, data, evaluation, risks, and appropriate use.
- Fairness and harmful bias management.
- Privacy and data-protection considerations.
- Safety, security, and misuse scenarios.
- Explainability and user communication.
- Accountability, transparency, and human oversight.
- Validating the system before and after release.
- Project documentation and presentation.
Practice: Complete the capstone, create a model or application card, document risks, and present both capabilities and limitations.
Practical exercise ideas
These activities help connect theory with practical AI problem-solving before the final project.
AI use-case analysis
Compare three possible AI use cases and define the task, inputs, outputs, users, and failure risks.
Review focus: clear objectives, measurable outcomes, and appropriate use of AI.
Route-planning agent
Represent a grid or network and compare search strategies for finding a path.
Review focus: state representation, cost, heuristics, and efficiency.
Prediction baseline
Prepare a dataset, train baseline models, compare metrics, and inspect incorrect predictions.
Review focus: preprocessing, leakage, validation, and generalization.
Text classification
Classify short text items and review where language ambiguity affects results.
Review focus: labels, representation, evaluation, and error analysis.
Document assistant
Design a grounded assistant for a small approved document collection.
Review focus: retrieval, citations, refusal behavior, and human review.
Risk review worksheet
Identify possible privacy, fairness, safety, security, and misuse concerns for an AI use case.
Review focus: affected users, mitigations, monitoring, and accountability.
Suggested twelve-week learning plan
This is an illustrative learning sequence. Confirm the institute's official timetable, session count, and delivery format before publishing it as a schedule.
| Week | Focus | Suggested milestone |
|---|---|---|
| 01 | AI fundamentals | Analyze use cases, goals, users, and failure risks. |
| 02 | Agents and problem solving | Model a search problem with states and actions. |
| 03 | Search and reasoning | Compare search or optimization approaches. |
| 04 | Data preparation | Clean and document a small dataset. |
| 05 | Supervised learning | Train and evaluate a baseline model. |
| 06 | Unsupervised learning | Explore grouping or dimensionality-reduction ideas. |
| 07 | NLP | Complete a small text-processing experiment. |
| 08 | Vision and neural networks | Plan or prototype an image-based task. |
| 09 | Generative AI | Design a grounded assistant or content workflow. |
| 10 | Evaluation and deployment | Create an evaluation report and deployment plan. |
| 11 | Responsible AI | Complete a risk and mitigation review. |
| 12 | Capstone presentation | Demonstrate the application, results, and limitations. |
Bring the topics together
Capstone project
Intelligent knowledge assistant
Build a practical AI application that helps users search and understand a controlled collection of documents. The project may use retrieval, text processing, classification, or a generative interface, depending on the selected scope.
Core project requirements
- Define a specific user problem and intended audience.
- Describe the system inputs, outputs, and limitations.
- Use a small approved document or dataset collection.
- Prepare and document the data used by the project.
- Implement a search, classification, or response workflow.
- Provide a way to review supporting information.
- Include tests for normal, invalid, and uncertain inputs.
- Record representative successful and failed examples.
- Document privacy, safety, bias, and misuse considerations.
- Present the architecture, evaluation, and next steps.
Evaluation requirements
- Define what counts as a useful or correct result.
- Use a small evaluation set separate from development examples.
- Review errors instead of reporting only one summary number.
- Describe where human review is required.
- Document data limitations and possible sources of bias.
- Explain how the system should be monitored after release.
Optional extensions
Add multilingual text support, feedback collection, role-based access, a simple API, a web interface, structured citations, model comparison, or monitoring dashboards. Expand the scope only after the core workflow is understandable and testable.
An AI prototype is not automatically safe or production-ready. Review data rights, privacy, security, evaluation quality, human oversight, and domain-specific risks before real-world use.
Suggested project structure
Keep application code, data descriptions, evaluation examples, documentation, and risk notes organized so another person can understand the project.
ai-project/
├── app/
│ ├── main.py
│ ├── pipeline.py
│ └── prompts/
├── data/
│ ├── README.md
│ └── sample-data/
├── evaluation/
│ ├── test-cases.json
│ └── results.md
├── docs/
│ ├── architecture.md
│ ├── limitations.md
│ └── risk-review.md
├── tests/
├── README.md
└── .gitignore
Do not commit private datasets, personal information, secret keys, confidential documents, or unapproved third-party content to a public repository.
AI evaluation and quality
AI quality depends on the task, data, users, evaluation method, and consequences of errors. A single metric rarely explains the complete system.
Does it solve the intended task?
Check whether outputs are useful for the defined objective and audience.
Does behavior remain consistent?
Test representative examples, repeated runs, edge cases, and changes in input conditions.
Is the data suitable?
Review completeness, accuracy, coverage, labels, provenance, and possible leakage.
Are harms distributed unevenly?
Compare performance and outcomes across relevant groups where appropriate.
What could go wrong?
Identify misuse, unsafe outputs, privacy risks, security threats, and harmful failure modes.
When should people review?
Define escalation, correction, and override processes for uncertain or high-impact results.
NIST describes trustworthy AI characteristics including validity and reliability, safety, security and resilience, accountability and transparency, explainability, privacy enhancement, and fairness with harmful bias managed. [48]
Tools and technologies
The exact tools may vary by project scope. The following technologies represent a practical learning toolkit for introductory AI application work.
- Python
- Jupyter Notebook
- NumPy
- pandas
- scikit-learn
- Matplotlib
- Git
- GitHub
- VS Code
- API concepts
Supporting concepts
- Data cleaning and exploratory analysis.
- Feature representation and preprocessing.
- Model training and evaluation.
- Text and image data handling.
- Application interfaces and deployment planning.
- Documentation, experiment tracking, and versioning.
Learning outcomes
By completing the proposed lessons and exercises, aim to demonstrate the following abilities:
- Explain major AI concepts and application areas.
- Frame a real-world problem as an AI task.
- Represent a simple problem for search or reasoning.
- Prepare structured data for a learning workflow.
- Train and compare basic models responsibly.
- Interpret metrics and inspect model errors.
- Describe NLP, computer vision, and neural-network concepts.
- Design a grounded generative-AI application workflow.
- Identify privacy, fairness, safety, and security risks.
- Document, evaluate, and present an AI project.
These are learning objectives, not guarantees of employment, certification, placement, or a particular AI role. Results depend on practice, mathematics, programming, project quality, and continued study.
Related career interests
This broad foundation can help you explore different AI and data-related directions.
- AI Developer
- Machine Learning Trainee
- Data Analyst
- Junior Data Scientist
- NLP Developer
- Computer Vision Trainee
- AI Application Developer
- Technology Research Assistant
Portfolio presentation ideas
- Explain the user problem and why AI is appropriate.
- Describe the data source, preparation, and limitations.
- Show a baseline and explain the evaluation method.
- Demonstrate successful and incorrect outputs.
- Discuss uncertainty and human-review requirements.
- Explain privacy, fairness, safety, and security risks.
- Present the architecture and possible deployment path.
- Describe what you would improve with more data or time.
Frequently asked questions
Who is this course for?
It is suitable for beginners who want a broad introduction to Artificial Intelligence as well as learners who want to compare AI specialization areas before choosing a deeper path.
Do I need programming experience?
Basic Python knowledge is recommended for practical coding exercises, but the course begins with conceptual AI topics and does not assume advanced programming experience.
What topics are covered?
The course covers AI fundamentals, intelligent agents, search, reasoning, data preparation, machine learning, NLP, computer vision, neural networks, generative AI, evaluation, deployment, and responsible AI.
Is this only a machine-learning course?
No. Machine learning is one important part of the curriculum. The course also introduces symbolic reasoning, search, language, vision, generative systems, evaluation, and governance.
Will I build a project?
The proposed capstone is an intelligent knowledge assistant or another scoped AI application. It should include documentation, evaluation examples, limitations, and a risk review.
Which tools are included?
The suggested toolkit includes Python, Jupyter, NumPy, pandas, scikit-learn, Matplotlib, Git, GitHub, VS Code, and API concepts. The exact tools may vary by project.
Does the course cover responsible AI?
Yes. The course includes fairness, privacy, safety, security, explainability, accountability, transparency, human oversight, and misuse analysis.
How long is the course?
The supplied course information proposes a duration of 12 weeks. Confirm the academy's official schedule before publishing or enrolling.
Can I start without mathematics?
You can begin with the conceptual modules. Basic percentages, averages, charts, and probability intuition will help with later data and evaluation topics. More advanced specialization may require deeper mathematics.
Can an AI project guarantee accurate results?
No. AI behavior depends on the task, data, model, evaluation method, environment, and user interaction. A project should document uncertainty, errors, and limitations.
How do I enroll?
This page is a frontend course-information demonstration. Enrollment, payment, scheduling, and admission workflows are not implemented here.
Does this course guarantee a job?
No. The course can support foundational learning and portfolio development, but it does not guarantee employment, placement, certification, or salary.
Explore intelligent systems
Build your AI learning path
Study AI foundations, experiment with data and models, evaluate results carefully, and design applications with responsible-use considerations.