Roadmap: Artificial Intelligence and Machine Learning¶

This roadmap explains how the AI/ML topic fits into broader learning paths, what to study before and after, and suggested timelines for different goals.

Table of Contents¶

• Where This Topic Fits
• Suggested Learning Paths
• Before You Start
• After You Finish
• Parallel Topics
• Timeline Estimates

Where This Topic Fits¶

AI/ML sits at the intersection of mathematics, software engineering, and domain expertise. It builds on Python programming, statistics, and linear algebra — and opens doors to research, data science, ML engineering, and generative AI development.

flowchart TD
    A["Python Basics\n[[async-python]]"] --> B["AI and Machine Learning\n(this topic)"]
    C["Statistics & Probability"] --> B
    D["Linear Algebra"] --> B
    E["SQL / Databases\n[[postgresql]]"] --> B
    B --> F["ML Engineering\nModule 09"]
    B --> G["LLMs & Generative AI\nModule 10"]
    B --> H["Research / PhD-track"]
    B --> I["Systems Architecture\n[[systems-architecture]]"]

The AI/ML topic requires Python proficiency and benefits from SQL knowledge. It leads into production ML engineering and cutting-edge LLM development.

Suggested Learning Paths¶

Path 1: ML Engineer (6–9 months)¶

Focus on getting production-ready quickly.

1. Module 01 — Introduction (orientation)
2. Module 02 — Math for ML (skim if comfortable with linear algebra)
3. Module 03 — Supervised Learning (core algorithms)
4. Module 05 — Model Evaluation (critical for production)
5. Module 07 — Deep Learning with PyTorch
6. Module 09 — ML Engineering (primary focus)
7. Module 12 — Capstone Project

Path 2: Data Scientist (4–6 months)¶

Focus on analysis, classical ML, and communication of results.

1. Modules 01–05 (all) — Foundations through evaluation
2. Module 06 — Neural Networks (conceptual understanding)
3. Module 11 — Responsible AI (bias, fairness, explainability)
4. Module 12 — Capstone Project

Path 3: Deep Learning Researcher (9–12 months)¶

Focus on depth across all modules.

1. All modules in order
2. Supplement with "Deep Learning" by Goodfellow/Bengio/Courville
3. Read original papers for each module's key algorithms

Path 4: LLM / Generative AI Developer (5–7 months)¶

Focus on Transformers, LLMs, and prompt engineering.

1. Modules 01–03 (foundations)
2. Module 05 (evaluation — essential for LLM eval)
3. Module 07 (deep learning basics)
4. Module 08 (NLP and Transformers — primary focus)
5. Module 10 (LLMs and Generative AI — primary focus)
6. Module 11 (Responsible AI)
7. Module 12 (Capstone)

Before You Start¶

Ensure you are comfortable with:

• Python — functions, classes, list comprehensions, numpy array operations, file I/O. See [[async-python]].
• Basic statistics — mean, variance, standard deviation, probability. Module 02 covers these formally but pre-exposure helps.
• Basic linear algebra — vectors, matrices, dot products. Again, Module 02 covers this — but familiarity speeds up comprehension.
• Command-line basics — creating virtual environments, installing packages with pip, running Python scripts.

After You Finish¶

Once you complete this topic, natural next steps include:

• [[systems-architecture]] — Understand distributed training, parameter servers, and inference at scale
• [[devops-platform-engineering]] — Container-based deployment, Kubernetes for ML workloads, GitOps
• [[postgresql]] — Advanced SQL for data pipelines, feature stores, and experiment tracking
• Original research papers — follow arXiv (cs.LG, cs.CL, stat.ML) for the latest advances
• Kaggle competitions — apply your skills on real datasets with community feedback
• Open-source contribution — PyTorch, HuggingFace, scikit-learn, and MLflow all welcome contributors

Parallel Topics¶

These topics can be studied alongside AI/ML:

• Statistics — probability theory, Bayesian inference, hypothesis testing
• Algorithms and Data Structures — efficient implementations of nearest-neighbor search, graph algorithms used in GNNs
• Cloud Computing — AWS SageMaker, GCP Vertex AI, Azure ML for managed training and deployment

Timeline Estimates¶