# AI and Machine Learning Fundamentals > An online, self-paced AI course from ARFA, the TIBLOGICS AI Academy. Intermediate level, about 20 hours, in English and French, with a verifiable certificate. Web page: https://tiblogics.com/learning-box/ai-ml-fundamentals ยท All tracks: https://tiblogics.com/learning-box.md ## Key facts - Level: Intermediate - Time: about 20 hours including hands-on work (6 modules, 27 lessons, 6 labs); self-paced, no deadline - Price: $487 one time for lifetime access to this track, or every track for $89 a month (cancel anytime) - Certificate: TIBLOGICS Certified: AI and Machine Learning Fundamentals, with a public verification page - Languages: English and French ## About Most people who work with AI have never been shown how it actually works. This track gives you a solid, vendor-neutral technical foundation without asking you to write code. You will learn how machines learn from data, what makes data fit for a model, how the common model types work in plain terms, and how to recognise vision, language, speech and document workloads. You will learn to read an evaluation properly: precision and recall, thresholds, error measures, and how to evaluate generative AI with test sets, rubrics and checked model judges. You will look inside foundation models (tokens, transformers, embeddings, retrieval-augmented generation, agents) and learn to apply them: choosing a model, deciding between prompting, retrieval and fine-tuning, understanding the main fine-tuning methods, inference settings and text-scoring measures, estimating cost, and launching with monitoring and a real human in the loop. The final module covers fairness, privacy, security and governance, including risk-based regulation and recognised frameworks. Skills this track builds also appear in entry-level cloud AI practitioner and AI fundamentals certifications. This track is independent: it is not affiliated with any vendor and is not official exam preparation. Lessons include small interactive demos you can edit, prompts you run on the page and Studio tools. Labs are done on the platform. The capstone is an AI solution proposal for a real problem in your own organisation, reviewed by a person. Who it is for: Professionals, analysts, managers and aspiring builders who want a solid technical foundation in how AI and machine learning work: enough to speak credibly with engineers, choose solutions, and prepare for entry-level cloud AI certifications. No coding required. ## What you will be able to do - Explain how machines learn, distinguish the main kinds of learning, and frame a business problem as a machine learning task, or recognise when it should not be one - Assess data for quality and bias, and explain splitting, overfitting and the common model types in plain terms - Choose and interpret evaluation metrics for classification, regression and generative AI, and tie them to business outcomes - Explain tokens, transformers, embeddings, retrieval-augmented generation, hallucination and agents - Choose a foundation model, decide between prompting, retrieval and fine-tuning, estimate running cost and plan a monitored launch with a human in the loop - Assess an AI system for fairness, privacy and security risks, and outline proportionate governance using risk tiers and recognised frameworks ## Curriculum ### Module 1: How Machines Learn Separate AI, machine learning and deep learning; understand supervised, unsupervised and reinforcement learning; learn the vocabulary of features, labels, training and inference; and follow the machine learning lifecycle from data to monitoring. - AI, machine learning and deep learning (24 min) - Supervised, unsupervised and reinforcement learning (25 min) - Features, labels, training and inference (24 min) - The machine learning lifecycle (25 min) - Module quiz ### Module 2: Data and Models Judge whether data is fit for a model, split it properly and recognise overfitting, understand the common model types in plain terms, and decide which approach suits a business problem, including when machine learning is the wrong answer. - Data quality and bias (25 min) - Splitting data and overfitting (25 min) - Common model types in plain terms (26 min) - Choosing the right approach, and when not to use ML (24 min) - Recognising AI workloads: vision, language, speech and documents (25 min) - Module quiz ### Module 3: Evaluating Models Read a confusion matrix and choose between accuracy, precision, recall and F1; understand thresholds, ROC and AUC, and regression error measures; evaluate generative AI with test sets, rubrics, human review and checked model judges; and keep model metrics tied to business outcomes. - Accuracy, precision, recall and the confusion matrix (27 min) - Thresholds, ROC curves and regression errors (25 min) - Evaluating generative AI (26 min) - Business metrics versus model metrics (23 min) - Module quiz ### Module 4: Generative AI and Foundation Models Understand tokens, context windows and transformers in plain language; the difference between pre-training, fine-tuning and prompting; embeddings and vector search; retrieval-augmented generation end to end; hallucination and grounding; multimodal models; and agents in brief. - Tokens, transformers and foundation models (26 min) - Embeddings and vector search (25 min) - Retrieval-augmented generation end to end (27 min) - Hallucination, grounding, multimodal models and agents (26 min) - Module quiz ### Module 5: Applying Foundation Models Choose a model on capability, cost, latency, context and data residency; decide between prompting, retrieval and fine-tuning; estimate running costs; and launch with evaluation, monitoring and a human in the loop designed as part of the whole system. - Choosing a model (25 min) - Prompting, RAG or fine-tuning: a decision guide (26 min) - Estimating cost (25 min) - Launching, monitoring and the human in the loop (27 min) - Customising and measuring a foundation model: methods, settings and scores (27 min) - Module quiz ### Module 6: Responsible, Secure and Governed AI Apply fairness, transparency and explainability; handle personal data lawfully and carefully; recognise prompt injection, data leakage, model theft and poisoning; understand shared responsibility with cloud providers; and use risk tiers and recognised frameworks to govern AI. - Fairness, transparency and explainability (26 min) - Privacy and data protection basics (24 min) - Security risks and shared responsibility (26 min) - Governance: risk tiers, frameworks and accountability (25 min) - Securing data and ML pipelines: the doors that matter (25 min) - Module quiz ## How you are assessed - A quiz after each module (80% to pass, retakes allowed). - A timed final exam: 35 questions in 60 minutes, 75% to pass, 90% for distinction, 3 attempts. - A capstone project reviewed by a person against a published rubric (70% to pass). - Each certificate has a public page at https://tiblogics.com/certificates/ that anyone can open to check it.