GenAI Foundations Navigator
A guided tour of the vocabulary everything else assumes: grounding, retrieval components, prompt structure, core parameters, and the point at which fine-tuning starts to make sense.
Interactive decision trees, graduate-level courses, and exam simulators for machine learning, data engineering, and generative AI practitioners. Built by Tarek Atwan.
A guided tour of the vocabulary everything else assumes: grounding, retrieval components, prompt structure, core parameters, and the point at which fine-tuning starts to make sense.
Work from your problem type and data characteristics to a shortlist of algorithms, with the trade-off that decided each branch stated in the open.
Encoding, scaling, outlier handling, feature creation and selection — branched separately for tabular, text, time-series and image data instead of averaged into one rule of thumb.
Navigate the landscape of evaluation metrics for classification, regression and ranking, including what to do when your classes are badly imbalanced.
A ten-module, vendor-neutral course running from source systems through ingestion, storage and modelling to decisions people actually trust.
Study guide for the AWS Certified AI Practitioner exam, with interactive flashcards, service-by-service coverage and quick-fire questions.
Timed practice exams with 189 questions across all five domains, matching the real AIF-C01 distribution, with an explanation behind every answer.
Diagrams covering confusion matrices, RAG pipelines and the complete LLM lifecycle, drawn for the AIF-C01 curriculum.
Thirteen interactive study tools and domain guides for the AWS Certified Generative AI Developer – Professional (AIP-C01) exam: deep study guides, 65 flashcards, gap cheatsheet, and 242 practice questions across timed simulators.
Interactive study guide for the Microsoft Fabric Data Engineering Associate exam: architecture, ingestion, Delta Lake, KQL, PySpark and the traps that catch people.
Score a candidate use case across six value dimensions and come out with one of four verdicts — build, buy, experiment, or kill — plus the rationale to defend it.
Work from a business problem to a solution class — rules, classical ML, deep learning, generative models, agents, or operations research — before anyone commits to a stack.
Choose the right LLM optimization for your constraint: RAG, fine-tuning, LoRA, quantization, distillation, and the cases where prompt engineering was already enough.
Pick metrics for foundation-model evaluation — perplexity, ROUGE, BERTScore, RAGAS, LLM-as-judge — and understand what each one quietly fails to measure.
A 19-module interactive course on Retrieval-Augmented Generation: 12 retrieval architectures (Classic, Hybrid, HyDE, Self-RAG, Corrective, Adaptive, GraphRAG, KAG, Agentic, Modular, Multimodal, Speculative), chunking, indexing, formal math, and a live playground.
All 15 weeks of graduate study notes from Georgia Tech OMSA Applied NLP: text preprocessing, classical ML, embeddings, RNNs/LSTMs, Transformers, BERT/GPT, sequence labeling, post-training alignment (RLHF/DPO), topic modeling, RAG, Agentic AI, multi-agent systems and GraphRAG — with 24 live interactive visualizers.
Practice for the Microsoft Power BI Data Analyst exam: 140 questions in the real exam formats with an explanation behind every answer, timed simulations, a study guide for every skill on the outline, weak-area tracking and a final review sheet.
No playbooks or internal nodes match your current filter. Try a broader query or clear filters.
Engineer clean data platforms and features, select the strongest model family, and align evaluation metrics to business cost.
Build core LLM fluency, choose between RAG, fine-tuning, and inference optimization, and evaluate reasoning & agentic behavior.
Combine structured domain guides with timed exam simulators across AWS AI Practitioner, AWS GenAI Pro, Microsoft Fabric, and Power BI.
These playbooks, decision trees, and courses grew out of real questions from workshops, graduate classrooms, and consulting engagements—the kind where the useful answer is knowing exactly what a technical decision depends on.
Collected and maintained by Tarek Atwan — author, educator, and data & AI consultant. Read essays at tarekatwan.com and explore consulting work at Ensemble Methods.