The Python Open Source Data Science Degree

Version 2.0 — 2026 Edition

A curated list of free or low-cost courses from reputable universities and organizations that satisfy the same requirements as an undergraduate Computer Science / Data Science degree, minus general education. Updated for 2026.

---

Curriculum Roadmap

1. Learning How to Learn
2. CS Basics & Programming
3. Math (Calculus → Linear Algebra → Statistics)
4. Databases
5. Data Science
6. Machine Learning
7. Large Language Models (LLMs)
8. Agentic AI & AI Engineering
9. MLOps & LLMOps
10. Computing Systems & CS Theory
11. Cloud, DevOps & Containers
12. Unix, Open Source & Version Control

---

Learning How to Efficiently Learn

If it has been a while since you were in the classroom, this is mandatory. These are high-leverage meta-skills that pay dividends across every other course.

Courses	School	Duration	Effort	Frequency	Prerequisites	Provider
Learning How to Learn	Deep Teaching Solutions	4 weeks	3-4 hrs/week	self-paced	none	Coursera
A Mind for Numbers (Book)	Barbara Oakley	—	—	—	none	Book

---

Computer Science Basics

Courses	School	Duration	Effort	Frequency	Prerequisites	Provider
CS50x: Introduction to Computer Science	Harvard	12 weeks	10-20 hrs/week	self-paced	none	edX
CS50P: Introduction to Programming with Python	Harvard	10 weeks	10-20 hrs/week	self-paced	none	edX
Mathematical Thinking in Computer Science	UC San Diego	6 weeks	2-5 hrs/week	once a month	none	Coursera

---

Programming

Courses	School	Duration	Effort	Frequency	Prerequisites	Provider
Python for Everybody Specialization	U of Michigan	8 months	3-6 hrs/week	self-paced	none	Coursera
Python OOP & Design Patterns	Duke	4 weeks	4-6 hrs/week	self-paced	Python basics	Coursera
Software Engineering: Introduction	UBC	6 weeks	6-8 hrs/week	self-paced	Programming basics	edX

---

Math

Single Variable Calculus

Courses	School	Duration	Effort	Frequency	Prerequisites	Provider
18.01x Single Variable Calculus	MIT	10 months	6-10 hrs/week	self-paced	pre-calculus	edX

Linear Algebra and Statistics

Courses	School	Duration	Effort	Frequency	Prerequisites	Provider
Linear Algebra - Foundations to Frontiers	UT Austin	15 weeks	6-10 hrs/week	self-paced	pre-calculus	edX
Introduction to Statistics	Stanford	7 weeks	5 hrs/week	self-paced	none	Coursera
Statistical Learning	Stanford	9 weeks	5 hrs/week	self-paced	Linear Algebra, Stats	edX

---

Databases

Courses	School	Duration	Effort	Frequency	Prerequisites	Provider
PostgreSQL for Everybody Specialization	U of Michigan	4 months	4-6 hrs/week	self-paced	Programming basics	Coursera
Databases: Relational Databases and SQL	Stanford	6 weeks	5-10 hrs/week	self-paced	none	edX
Vector Databases: from Embeddings to Applications	Weaviate / DeepLearning.AI	1 week	2-3 hrs	self-paced	Python basics	DeepLearning.AI

---

Data Science

Courses	School	Duration	Effort	Frequency	Prerequisites	Provider
Applied Data Science with Python Specialization	U of Michigan	5 months	7 hrs/week	self-paced	Python	Coursera
Data Science: Visualization	Harvard	8 weeks	2-4 hrs/week	self-paced	R basics	edX
Practical Data Ethics	fast.ai	4 weeks	4 hrs/week	self-paced	none	fast.ai

---

Machine Learning

This section is a core requirement. It covers classical and modern machine learning, deep learning, and the foundational skills needed to build, evaluate, and deploy models.

Courses

Courses	School	Duration	Effort	Frequency	Prerequisites	Provider
Machine Learning Specialization	Stanford / DeepLearning.AI	3 months	9 hrs/week	self-paced	Python, Linear Algebra	Coursera
Practical Deep Learning for Coders	fast.ai	10 weeks	8-10 hrs/week	self-paced	Python, some math	fast.ai
Deep Learning Specialization	DeepLearning.AI	5 months	4-5 hrs/week	self-paced	ML basics	Coursera

Topics Covered

• Supervised Learning
• Unsupervised Learning
• Regression
• Classification
• Decision Trees
• Ensemble Methods
• Support Vector Machines
• Clustering
• Dimensionality Reduction
• Feature Engineering
• Neural Networks
• Convolutional Neural Networks
• Recurrent Neural Networks
• Deep Learning
• Model Evaluation
• Hyperparameter Optimization
• Explainable AI
• Statistical Learning

Key Libraries and Frameworks

• NumPy
• pandas
• SciPy
• scikit-learn
• PyTorch
• TensorFlow
• Keras
• XGBoost
• LightGBM

---

Large Language Models (LLMs)

This section covers the theory, architecture, and practical application of large language models. Topics range from transformer fundamentals and prompt engineering to fine-tuning, retrieval-augmented generation, and responsible deployment.

Courses

Courses	School	Duration	Effort	Frequency	Prerequisites	Provider
Hugging Face NLP Course	Hugging Face	6 weeks	4-6 hrs/week	self-paced	Python, ML basics	Hugging Face
CS324: Large Language Models	Stanford	10 weeks	6-8 hrs/week	self-paced	Deep Learning	Stanford
LLM Bootcamp	Full Stack Deep Learning	8 hours	self-paced	self-paced	ML basics	FSDL
LLM University	Cohere	~20 hours	self-paced	self-paced	Python	Cohere
DeepLearning.AI Short Courses	DeepLearning.AI	1-3 hrs each	1-3 hrs	self-paced	Python	DeepLearning.AI

Topics Covered

• Transformer Architecture
• Self-Attention
• Tokenization
• Embeddings
• Prompt Engineering
• Structured Outputs
• Function Calling
• Context Windows
• Retrieval-Augmented Generation
• Fine-Tuning
• LoRA and QLoRA
• Open-Weight Models
• Quantization
• Inference Optimization
• LLM Evaluation
• Safety and Alignment
• Multimodal Models

Key Libraries and Frameworks

• Hugging Face Transformers
• Tokenizers
• Datasets
• PEFT
• TRL
• llama.cpp
• Ollama
• vLLM
• OpenAI SDK
• Anthropic SDK
• Google GenAI SDK

---

Agentic AI & AI Engineering

This section covers the design, construction, and deployment of AI agents and multi-agent systems. It includes tool use, memory, planning, and orchestration frameworks that form the foundation of modern AI engineering.

Courses

Courses	School	Duration	Effort	Frequency	Prerequisites	Provider
AI Agents	DeepLearning.AI	2 hrs	2 hrs	self-paced	LLM basics	DeepLearning.AI
Building AI Browser Agents	DeepLearning.AI	2-3 hrs	2-3 hrs	self-paced	LLM basics	DeepLearning.AI
OpenAI Agents SDK	OpenAI	3-6 hrs	self-paced	self-paced	Python	OpenAI
Model Context Protocol	Anthropic	3-4 hrs	self-paced	self-paced	Python	MCP Documentation
LangGraph Academy	LangChain	8-15 hrs	self-paced	self-paced	Python, LLM basics	LangChain
CrewAI Documentation	CrewAI	4-6 hrs	self-paced	self-paced	Python	CrewAI
AutoGen Documentation	Microsoft	6-10 hrs	self-paced	self-paced	Python	Microsoft
AG2 Documentation	AG2	self-paced	self-paced	self-paced	Python	AG2
OpenHands Documentation	OpenHands	4-8 hrs	self-paced	self-paced	Python	OpenHands

Topics Covered

• AI Agents
• Tool Calling
• Function Calling
• Agent Memory
• State Management
• Planning
• Reflection
• Multi-Agent Systems
• Agent Communication
• Browser Agents
• Coding Agents
• Retrieval-Augmented Generation
• Model Context Protocol
• Human-in-the-Loop Workflows
• Long-Running Agents
• Agent Evaluation
• Agent Safety
• Workflow Orchestration

Key Libraries and Frameworks

• OpenAI Agents SDK
• LangGraph
• LangChain
• CrewAI
• AutoGen
• AG2
• OpenHands
• LlamaIndex
• DSPy
• PydanticAI
• Haystack

Capstone Project Ideas

• Research Assistant
• Enterprise RAG Chatbot
• AI Coding Assistant
• Browser Automation Agent
• Multi-Agent Software Development Team
• Document Processing Pipeline
• Customer Support Agent
• Personal Knowledge Management Assistant

---

MLOps & LLMOps

This section covers the operational side of machine learning and LLM systems — from experiment tracking and model registries to prompt versioning, evaluation, observability, and production deployment. MLOps and LLMOps are now essential skills for any practitioner deploying AI in the real world.

Courses

Courses	School	Duration	Effort	Frequency	Prerequisites	Provider
MLOps Specialization	DeepLearning.AI	4 months	4 hrs/week	self-paced	ML basics	Coursera
Weights & Biases Courses	Weights & Biases	varies	self-paced	self-paced	Python, ML	W&B

Topics Covered

• Model Deployment
• Experiment Tracking
• Model Registries
• Prompt Versioning
• LLM Evaluation
• Agent Evaluation
• Observability
• Tracing
• Cost Monitoring
• Model Serving
• Data and Model Drift
• Safety Monitoring
• CI/CD for Machine Learning
• LLM Application Testing

---

Computing Systems

Courses	School	Duration	Effort	Frequency	Prerequisites	Provider
Build a Modern Computer from First Principles I	Hebrew University	6 weeks	10-15 hrs/week	self-paced	none	Coursera
Build a Modern Computer from First Principles II	Hebrew University	6 weeks	10-15 hrs/week	self-paced	Part I	Coursera
Introduction to Operating Systems	Georgia Tech	8 weeks	5-8 hrs/week	self-paced	Part II	Udacity

---

Computer Science Theory

Courses	School	Duration	Effort	Frequency	Prerequisites	Provider
Algorithms, Part I	Princeton	6 weeks	6-12 hrs/week	once a month	Programming basics	Coursera
Algorithms, Part II	Princeton	6 weeks	6-12 hrs/week	once a month	Part I	Coursera
Automata Theory	Stanford	6 weeks	4-8 hrs/week	self-paced	Theory basics	Coursera

---

Cloud Computing

Cloud is now a foundational skill for data scientists and engineers.

Courses	School	Duration	Effort	Frequency	Prerequisites	Provider
Cloud Computing Foundations	Duke	5 weeks	4-6 hrs/week	self-paced	Linux basics	Coursera
AWS Cloud Practitioner Essentials	AWS	6 hours	self-paced	—	none	Coursera
Google Cloud Fundamentals: Core Infrastructure	Google	1 week	8-10 hrs	self-paced	none	Coursera

Recommended Certification Path

Pursue one cloud provider certification (AWS, GCP, or Azure) at the associate level after completing the foundational course.

---

DevOps and Containers

Courses	School	Duration	Effort	Frequency	Prerequisites	Provider
Docker for Beginners	Docker	4-8 hrs	self-paced	—	Linux basics	Docker
Kubernetes for Developers	Linux Foundation	14 weeks	2-3 hrs/week	self-paced	Docker basics	edX
DevOps, DataOps, MLOps	Duke	7 weeks	4-6 hrs/week	self-paced	Programming basics	Coursera

---

Computer Applications

Courses	School	Duration	Effort	Frequency	Prerequisites	Provider
Cryptography I	Stanford	7 weeks	5 hrs/week	once a month	Linear Algebra	Coursera
Computer Networking: A Top-Down Approach	UMass	8 weeks	5 hrs/week	self-paced	CS basics	Free (textbook)

---

Unix and Linux

Courses	School	Duration	Effort	Frequency	Prerequisites	Provider
The Unix Workbench	JHU	4 weeks	4 hrs/week	once a month	none	Coursera
Linux Command Line Basics	Udacity	1 week	5 hrs/week	self-paced	none	Udacity

---

Open Source Development and Version Control

Courses	School	Duration	Effort	Frequency	Prerequisites	Provider
Version Control with Git	Google	1 week	5 hrs	self-paced	none	Coursera
Open Source Software Development, Linux and Git Specialization	Linux Foundation	4 months	3-5 hrs/week	self-paced	some programming	Coursera
GitHub Actions	GitHub	1-2 hrs	self-paced	—	Git basics	GitHub Skills

---

Recommended Learning Order

1. Learning How to Learn
2. CS50x → CS50P
3. Python for Everybody
4. Math (Calculus → Linear Algebra → Statistics)
5. Algorithms Part I & II
6. Databases
7. Data Science
8. Computing Systems
9. Machine Learning Specialization
10. Practical Deep Learning for Coders
11. Deep Learning Specialization
12. Large Language Models (LLMs)
13. Agentic AI & AI Engineering
14. Cloud Computing + DevOps
15. MLOps & LLMOps

---

Changelog

Version 2.0 — 2026 Edition

• Added: ## Curriculum Roadmap section near the top for navigation
• Renamed: Machine Learning and AI → ## Machine Learning with topics and libraries subsections
• Removed: Old Large Language Models and Generative AI subsection from inside Machine Learning
• Added: New top-level ## Large Language Models (LLMs) section with 5 courses, topics, and libraries
• Added: New top-level ## Agentic AI & AI Engineering section with 9 courses, topics, libraries, and capstone ideas
• Renamed: MLOps subsection → top-level ## MLOps & LLMOps with expanded topics list
• Updated: Recommended Learning Order to reflect the new 15-step curriculum path
• Preserved: All existing course links and foundational sections unchanged