Timeline

Visual overview of the 18-month curriculum

0% Time Elapsed
566 Days Remaining
Dec 11 Curriculum Progress Jun 30
Today

Phase Timeline

01/26
02/26
03/26
04/26
05/26
06/26
07/26
08/26
09/26
10/26
11/26
12/26
01/27
02/27
03/27
04/27
05/27
06/27
1 The Mathematical Substrate
Phase 1 Upcoming
The Mathematical Substrate
Dates Dec 11 – Mar 15
Focus Rigorous Proofs & Geometry
Goal Ability to derive gradients and visualize high-dimensional spaces.
Checkpoint Linear Algebra Done Right (Axler)
2 CS Fundamentals & Systems
Phase 2 Upcoming
CS Fundamentals & Systems
Dates Mar 16 – May 30
Focus Algorithms & Systems
Goal Optimization of compute and memory; writing efficient kernels.
Checkpoint The Algorithm Design Manual (Skiena)
3 Classical ML Theory
Phase 3 Upcoming
Classical ML Theory
Dates Jun 1 – Aug 15
Focus Statistical Learning
Goal Understanding bias, variance, and generalization bounds.
Checkpoint Pattern Recognition & ML (Bishop)
4 Deep Learning
Phase 4 Upcoming
Deep Learning
Dates Aug 16 – Nov 30
Focus Architectures (Transformers)
Goal Intuitive grasp of modern layers, attention, and normalization.
Checkpoint Understanding Deep Learning (Prince)
5 Frontier Systems
Phase 5 Upcoming
Frontier Systems
Dates Dec 1 – Feb 28
Focus Distributed Training (CUDA)
Goal Training models beyond single-GPU limits; scaling laws.
Checkpoint CMU 10-714 (Needle)
6 Frontier Research Topics
Phase 6 Upcoming
Frontier Research Topics
Dates Mar 1 – May 15
Focus LLMs, Generative AI
Goal Implementing papers from scratch; contributing to open science.
Checkpoint Stanford CS324 (LLMs)
7 Research & Portfolio
Phase 7 Upcoming
Research & Portfolio
Dates May 16 – Jun 30
Focus Reproduction & Innovation
Goal Implementing papers from scratch; contributing to open science.
Checkpoint Reproduction Project
Assignment 1.1 Upcoming
The Operator-Theoretic View of Neural Networks
Phase Phase 1
Type Mathematical Proofs & Conceptual Synthesis
Start Jan 10
Deadline Feb 7
Checkpoint Mastery of Chapters 1–7 of Linear Algebra Done Right (Axler). You must be comfortable with the concept of a linear map independent of its matrix representation.
Deliverables A LaTeX-typeset problem set (PDF), A companion blog post visualization
Assignment 1.2 Upcoming
The Calculus of Optimization (Autodiff Foundations)
Phase Phase 1
Type Mathematical Derivation & Implementation
Start Feb 1
Deadline Mar 15
Checkpoint Completion of Hubbard & Hubbard (Chapters on Derivatives as Linear Maps/Jacobians).
Deliverables Python Library (minigrad_tensor.py), Derivations
Assignment 2.1 Upcoming
The Computational Graph Optimizer
Phase Phase 2
Type Algorithmic Implementation (C++ preferred, Python accepted)
Start Apr 1
Deadline Apr 30
Checkpoint The Algorithm Design Manual (Skiena), Graph Algorithms chapters.
Deliverables A CLI tool for graph scheduling, Blog Output: "Graph Theory in the VRAM."
Assignment 2.2 Upcoming
The "Strided Memory" Challenge (Systems)
Phase Phase 2
Type Systems Programming (C/C++)
Start May 1
Deadline May 30
Checkpoint CS:APP (Bryant & O'Hallaron), chapters on Memory Hierarchy/Cache.
Deliverables Benchmarking Report, GitHub Repository
Assignment 3.1 Upcoming
Kernel Methods and Dual Representations
Phase Phase 3
Type Theoretical Derivation & Implementation
Start Jun 15
Deadline Jul 15
Checkpoint Pattern Recognition and Machine Learning (Bishop) Ch 6 (Kernel Methods).
Deliverables Theoretical derivations of the dual formulation, Python implementation of Kernel Ridge Regression
Assignment 3.2 Upcoming
Latent Variable Models (GMMs)
Phase Phase 3
Type Algorithm Implementation (from scratch)
Start Jul 15
Deadline Aug 15
Checkpoint Bishop Ch 9 (Mixture Models & EM).
Deliverables Blog Post: "The Expectation-Maximization Algorithm Visualized."
Assignment 4.1 Upcoming
"MicroGrad++" – The Autograd Engine
Phase Phase 4
Type Software Engineering / Library Building
Start Aug 16
Deadline Oct 15
Checkpoint Understanding Deep Learning (Prince) Ch 7; Karpathy's "Micrograd".
Deliverables A Python library on GitHub
Assignment 4.2 Upcoming
The Transformer from Scratch
Phase Phase 4
Type Model Implementation
Start Oct 16
Deadline Nov 30
Checkpoint Stanford CS25, "Attention Is All You Need" paper.
Deliverables PyTorch Implementation and Training Report
Assignment 5.1 Upcoming
High-Performance Kernels (CUDA/Triton)
Phase Phase 5
Type GPU Programming
Start Dec 1
Deadline Jan 30
Checkpoint GPU Mode videos; Programming Massively Parallel Processors.
Deliverables Triton kernel implementation, Benchmarking report comparing performance against PyTorch
Assignment 5.2 Upcoming
Distributed Training Simulation
Phase Phase 5
Type Systems Simulation
Start Feb 1
Deadline Feb 28
Checkpoint CMU 10-714; Papers on ZeRO/Megatron.
Deliverables Distributed training simulation code, Report on communication overhead analysis
Assignment 6.1 Upcoming
Alignment (RLHF / DPO)
Phase Phase 6
Type Model Fine-Tuning
Start Mar 1
Deadline Apr 15
Checkpoint Stanford CS324; Spinning Up in Deep RL.
Deliverables DPO implementation code, Qualitative analysis of aligned model behavior
Assignment 6.2 Upcoming
Diffusion Models (Generative AI)
Phase Phase 6
Type Implementation
Start Apr 16
Deadline May 15
Checkpoint Understanding Deep Learning Ch 18.
Deliverables Blog Post
Assignment 7 Upcoming
The Paper Reproduction
Phase Phase 7
Type Research Implementation
Start May 16
Deadline Jun 30
Checkpoint Mastery of the "3-Pass Reading Approach".
Deliverables Clean, documented repository, Ablation study, High-quality blog post
Completed Current Upcoming Assignment Deadline

Phase Details

1
Upcoming

The Mathematical Substrate

Focus: Rigorous Proofs & Geometry
Dates: Dec 11 – Mar 15
Checkpoint: Linear Algebra Done Right (Axler)
Assignments:
1.1 1.2
2
Upcoming

CS Fundamentals & Systems

Focus: Algorithms & Systems
Dates: Mar 16 – May 30
Checkpoint: The Algorithm Design Manual (Skiena)
Assignments:
2.1 2.2
3
Upcoming

Classical ML Theory

Focus: Statistical Learning
Dates: Jun 1 – Aug 15
Checkpoint: Pattern Recognition & ML (Bishop)
Assignments:
3.1 3.2
4
Upcoming

Deep Learning

Focus: Architectures (Transformers)
Dates: Aug 16 – Nov 30
Checkpoint: Understanding Deep Learning (Prince)
Assignments:
4.1 4.2
5
Upcoming

Frontier Systems

Focus: Distributed Training (CUDA)
Dates: Dec 1 – Feb 28
Checkpoint: CMU 10-714 (Needle)
Assignments:
5.1 5.2
6
Upcoming

Frontier Research Topics

Focus: LLMs, Generative AI
Dates: Mar 1 – May 15
Checkpoint: Stanford CS324 (LLMs)
Assignments:
6.1 6.2
7
Upcoming

Research & Portfolio

Focus: Reproduction & Innovation
Dates: May 16 – Jun 30
Checkpoint: Reproduction Project
Assignments:
7

Assignment Schedule

ID Assignment Phase Start Deadline Status
1.1 The Operator-Theoretic View of Neural Networks 1 Jan 10 Feb 7 Upcoming
1.2 The Calculus of Optimization (Autodiff Foundations) 1 Feb 1 Mar 15 Upcoming
2.1 The Computational Graph Optimizer 2 Apr 1 Apr 30 Upcoming
2.2 The "Strided Memory" Challenge (Systems) 2 May 1 May 30 Upcoming
3.1 Kernel Methods and Dual Representations 3 Jun 15 Jul 15 Upcoming
3.2 Latent Variable Models (GMMs) 3 Jul 15 Aug 15 Upcoming
4.1 "MicroGrad++" – The Autograd Engine 4 Aug 16 Oct 15 Upcoming
4.2 The Transformer from Scratch 4 Oct 16 Nov 30 Upcoming
5.1 High-Performance Kernels (CUDA/Triton) 5 Dec 1 Jan 30 Upcoming
5.2 Distributed Training Simulation 5 Feb 1 Feb 28 Upcoming
6.1 Alignment (RLHF / DPO) 6 Mar 1 Apr 15 Upcoming
6.2 Diffusion Models (Generative AI) 6 Apr 16 May 15 Upcoming
7 The Paper Reproduction 7 May 16 Jun 30 Upcoming