Hitesh Sahu
Hitesh SahuHitesh Sahu
  1. Home
  2. β€Ί
  3. posts
  4. β€Ί

  5. …

  6. β€Ί

  7. 0 ML Scientist Roadmap

Loading ⏳
Fetching content, this won’t take long…


πŸ’‘ Did you know?

πŸ¦₯ Sloths can hold their breath longer than dolphins 🐬.

πŸͺ This website uses cookies

No personal data is stored on our servers however third party tools Google Analytics cookies to measure traffic and improve your website experience. Learn more

Loading ⏳
Fetching content, this won’t take long…


πŸ’‘ Did you know?

🍯 Honey never spoils β€” archaeologists found 3,000-year-old jars still edible.
AI-Machine-Learning
    AI-AgenticAI
    AI-DeepLearning
    AI-GenAI
    AI-Infrastructure
    AI-Machine-Learning

    • Machine Learning Learning Path

    • Stanford AI Scientist Roadmap 2026

    • Machine Learning: Introduction and Core Algorithms

    • Linear Regression Explained: Single Variable and Multivariate Models with Gradient Descent

    • Evaluating a Hypothesis in Neural Networks

    • Bias-Variance Dilemma

    • Cost Function Regularization: Balancing Bias and Variance in Machine Learning Models

    • Polynomial Regression

    • Normal Equation in Linear Regression: Formula, Intuition, and Comparison with Gradient Descent

    • Logistic Regression for Classification: Concept, Sigmoid Function, Cost Function, and Implementation

    • Logistic Regression for Classification: Concept, Sigmoid Function, Cost Function, and Implementation

    • Support Vector Machines (SVM): Maximizing Margins for Robust Machine Learning Models

    • XGBoost (Extreme Gradient Boosting) Explained

    • Dimensionality Reduction in Machine Learning

    • Principal Component Analysis (PCA) Explained

    • t-SNE (t-distributed Stochastic Neighbor Embedding) Explained

    • K-Means Clustering

    • Anomaly Detection: Identifying Rare and Unusual Patterns in Data

    • Anomaly Detection Using Gaussian Distribution in Machine Learning

    • Anomaly Detection Using Multivariate Gaussian Distribution

    • Recommender Systems: Collaborative Filtering, Content-Based Filtering, and Hybrid Approaches

    • Collaborative Filtering: Building Recommender Systems with Feature Learning

    • Photo OCR: Sliding Window Detection, Character Segmentation and Recognition

    • Large Scale Machine Learning: Training Models on Massive Datasets

    • Stochastic Gradient Descent (SGD): Efficient Optimization for Large Datasets

    • MapReduce for Large-Scale Machine Learning: Distributed Training at Scale

    • AI-Machine-Learning Index

    AI-Math
    AWS
    Azure
    kubernetes
    Management
    Programming
    Terraform
    Z_Appendix
Cover Image for Stanford AI Scientist Roadmap 2026
AI-Machine-Learning

Stanford AI Scientist Roadmap 2026

A complete self-study roadmap built entirely from Stanford University's publicly available AI courses. Learn mathematics, machine learning, deep learning, reinforcement learning, large language models, AI systems, RAG, agentic AI, and production deployment through a structured path from foundations to real-world AI engineering.

Stanford
Artificial Intelligence
AI Engineering
Machine Learning
Deep Learning
Reinforcement Learning
← Previous

GPU Autoscaling on Kubernetes: KEDA, HPA, and Cluster Autoscaler

Next β†’

Deploying Agents at Scale

πŸŽ“ Stanford AI/ML Scientist Roadmap 2026

πŸ“ Mathematics β†’ πŸ€– Machine Learning β†’ 🧠 Deep Learning β†’ πŸ’¬ LLMs β†’ 🏭 Production AI Systems

A free, self-study path built entirely from Stanford's publicly available courses. Every course lists its ID, what it covers, where to find the slides, and a direct link to the official lecture videos where one exists.

Stanford AI/ML scientist roadmap from math foundations to production systems

πŸ“Œ How to use this: Watch lectures + do the public assignments. ~70% depth per stage is enough to advance. Whole path β‰ˆ 9–14 months part-time. Follow the main sequence; use the branches by interest.

Stage 0 β€” Mathematics & Foundations πŸ“

Build the mathematical foundation every AI engineer needs.

Stage 0 roadmap covering linear algebra, probability, and convex optimization courses

IDCourseCoversLecture SlidesPlaylist
MATH51LINEAR ALGEBRA & MULTIVARIABLE CALCULUSVectors, matrices, transformations, gradients, integralsSiteNo public video β€” use MIT 18.06 (Strang)
CS109PROBABILITY FOR COMPUTER SCIENTISTSProbability, random variables, distributions, Bayes, inferenceSite▢️ YouTube (2022)
EE364ACONVEX OPTIMIZATION I (BOYD)Convex sets/functions, optimization, duality, gradient methodsSlides▢️ YouTube (2023)

Stage 1 β€” Machine Learning πŸ€–

The classical ML core that underpins modern AI.

Stage 1 roadmap covering classical machine learning courses like CS229

IDCourseCoversLecture SlidesPlaylist
CS221ARTIFICIAL INTELLIGENCESearch, reasoning, logic, planning, decision-makingSite▢️ YouTube (2019)
CS229MACHINE LEARNINGSupervised & unsupervised learning, statistical learning, RL basicsNotes▢️ YouTube (2022)
CS229MMACHINE LEARNING THEORYGeneralization, VC dimension, PAC learning, bounds, guaranteesSiteNo public video β€” notes only
CS228PROBABILISTIC GRAPHICAL MODELSBayesian networks, Markov models, inference, structure learningSiteNo Stanford playlist β€” Koller PGM (Coursera)

Stage 2 β€” Deep Learning & Decision Making 🧠

Neural networks, perception, reinforcement learning, and decision making.

Stage 2 roadmap covering deep learning, computer vision, and reinforcement learning courses

IDCourseCoversLecture SlidesPlaylist
CS230DEEP LEARNINGNeural nets, backprop, optimization, regularization, training strategyLectures▢️ YouTube (2018)
CS231nDEEP LEARNING FOR COMPUTER VISIONCNNs, vision transformers, detection, segmentation, generative modelsSite▢️ YouTube (2025)
CS234REINFORCEMENT LEARNINGMDPs, value/policy learning, deep RL, explorationSite▢️ YouTube (2024)
CS238DECISION MAKING UNDER UNCERTAINTYMDPs, POMDPs, decision theory, estimation, controlSiteNo public video β€” free textbook + slides

Advanced sequential decision-making continues in CS239 (AA229) if you want to go deeper.

Stage 3 β€” LLMs & Generative AI πŸ’¬

From NLP fundamentals to building and training large language models.

Stage 3 roadmap covering NLP, transformers, and large language model courses

IDCourseCoversLecture SlidesPlaylist
CS224NNLP WITH DEEP LEARNINGWord embeddings, attention, transformers, LLM foundationsSite▢️ YouTube (2024)
CS224UNATURAL LANGUAGE UNDERSTANDINGSemantics, dialogue systems, meaning & groundingSite▢️ YouTube (2023)
CS25TRANSFORMERS UNITEDFoundation models, transformer architectures, multimodal, scalingSite▢️ YouTube
CS324LARGE LANGUAGE MODELSLLM theory, scaling, capabilities, harms, evaluationNotes & readingsNo public video β€” notes only
CS336 ⭐LANGUAGE MODELING FROM SCRATCHTokenization, transformer impl, GPUs/Triton, parallelism, inference, scaling lawsSite Β· Repo▢️ YouTube (2025)

Stage 4 β€” AI Systems & Production 🏭

Build, deploy, and operate AI systems in the real world.

Stage 4 roadmap covering MLOps and trustworthy ML systems courses

IDCourseCoversLecture SlidesPlaylist
CS329SMACHINE LEARNING SYSTEMS DESIGNMLOps, model serving, monitoring, A/B testing, evaluation, scalingSyllabusNo full playlist β€” Demo Day + guest tutorials
CS329TTRUSTWORTHY ML: LLMS & APPLICATIONSRAG, agentic AI, evaluation, reliability, responsible deploymentSiteNo public video β€” slides only

Note: CS329S last ran Winter 2022 β€” strong on classic MLOps, pre-LLM. CS329T is the modern complement: it's the only course on this roadmap that teaches RAG and agentic AI directly. For anything beyond the courses ( fine-tuning workflows, vLLM/Triton serving, vector DBs), Chip Huyen's AI Engineering (O'Reilly, 2024) is the current reference.

🦾 Optional β€” Robotics Track

For autonomy / robotics roles. Tangential to a generic AI-engineering path. Both are slides-only (no public video).

Optional robotics track roadmap covering robot autonomy courses

IDCourseCoversLecture SlidesPlaylist
CS237APRINCIPLES OF ROBOT AUTONOMY IPerception, localization, planning, controlSiteNo public video β€” slides only
CS237BPRINCIPLES OF ROBOT AUTONOMY IIRL, manipulation, imitation learning, human intent inferenceSiteNo public video β€” slides only

πŸ—ΊοΈ Suggested Sequence

Main path: Math β†’ ML β†’ Deep Learning β†’ LLMs β†’ Production

MATH51 / CS109 / EE364A
        β”‚
      CS229 ──► CS229M ──► CS228
        β”‚
      CS230 ──┬── CS231n         (Vision)
              β”œβ”€β”€ CS234 ──► CS238 (RL / decision-making)
              └── CS224U          (Language)
        β”‚
   converge β–Ί  CS224N ──► CS25 ──► CS324 ──► CS336
        β”‚
      CS329S ──► CS329T          (Production)

Optional branch:  CS237A ──► CS237B   (Robotics)

πŸ”— Master Link Reference

IDCourseSiteVideo
MATH51LINEAR ALGEBRA & CALCULUShttps://web.stanford.edu/class/math51/No public video β€” use MIT 18.06 (Strang)
CS109PROBABILITYhttps://web.stanford.edu/class/cs109/▢️ YouTube (2022)
EE364ACONVEX OPTIMIZATIONhttps://ee364a.stanford.edu/▢️ YouTube (2023)
CS221ARTIFICIAL INTELLIGENCEhttps://web.stanford.edu/class/cs221/▢️ YouTube (2019)
CS229MACHINE LEARNINGhttps://cs229.stanford.edu/▢️ YouTube (2022)
CS229MMACHINE LEARNING THEORYhttps://cs229m.stanford.edu/No public video β€” notes only
CS228PROBABILISTIC GRAPHICAL MODELShttps://web.stanford.edu/class/cs228/No Stanford playlist β€” Koller PGM (Coursera)
CS230DEEP LEARNINGhttps://cs230.stanford.edu/▢️ YouTube (2018)
CS231nDEEP LEARNING FOR CVhttps://cs231n.stanford.edu/▢️ YouTube (2025)
CS234REINFORCEMENT LEARNINGhttps://web.stanford.edu/class/cs234/▢️ YouTube (2024)
CS238DECISION MAKING UNDER UNCERTAINTYhttps://web.stanford.edu/class/cs238/No public video β€” free textbook + slides
CS224NNLP WITH DEEP LEARNINGhttps://web.stanford.edu/class/cs224n/▢️ YouTube (2024)
CS224UNATURAL LANGUAGE UNDERSTANDINGhttps://web.stanford.edu/class/cs224u/▢️ YouTube (2023)
CS25TRANSFORMERS UNITEDhttps://web.stanford.edu/class/cs25/▢️ YouTube
CS324LARGE LANGUAGE MODELShttps://stanford-cs324.github.io/No public video β€” notes only
CS336LM FROM SCRATCHhttps://cs336.stanford.edu/▢️ YouTube (2025)
CS329SML SYSTEMS DESIGNhttps://stanford-cs329s.github.io/No full playlist β€” Demo Day + guest tutorials
CS329TTRUSTWORTHY ML: LLMS & APPShttps://web.stanford.edu/class/cs329t/No public video β€” slides only
CS237APRINCIPLES OF ROBOT AUTONOMY Ihttps://web.stanford.edu/class/cs237a/No public video β€” slides only
CS237BPRINCIPLES OF ROBOT AUTONOMY IIhttps://web.stanford.edu/class/cs237b/No public video β€” slides only

▢️ = full official lecture playlist, verified. 11 of these 20 courses have public Stanford video; the other 9 were never recorded publicly (slides/notes only, or Canvas-restricted) β€” those rows point to the best legitimate substitute instead of a fake link.

✨ All courses are free to audit via lecture videos or notes. Enrolled/certificate versions are available through Stanford Online and the Stanford AI Professional Program. 🎯 Follow the sequence. Build projects. Master AI engineering.

Related Posts

  • Machine Learning Learning Path β€” the topic-by-topic study path that fills in this roadmap
  • Linear Algebra for Machine Learning β€” the math foundations this roadmap assumes
Hitesh Sahu
Written by Hitesh Sahu, a passionate developer and blogger.

Sat Jun 20 2026

Share This on

← Previous

GPU Autoscaling on Kubernetes: KEDA, HPA, and Cluster Autoscaler

Next β†’

Deploying Agents at Scale

AI-Machine-Learning/0-ML-Scientist-Roadmap
Let's work together
hiteshkrsahu@gmail.com
Munich πŸ₯¨, Germany πŸ‡©πŸ‡ͺ, EU
Playstore
Hitesh Sahu's apps on Google Play Store
Need Help?
Let's Connect
Navigation
Β  Home/About
Β  Skills
Β  Work/Projects
Β  Lab/Experiments
Β  Contribution
Β  Awards
Β  Art/Sketches
Β  Thoughts
Β  Contact
Links
Β  Sitemap
Β  Legal Notice
Β  Privacy Policy

Made with

NextJS logo

NextJS by

hitesh Sahu

| Β© 2026 All rights reserved.