Personal Curriculum
Learning tracks — in parallel with FAIRE
Ten tracks, running simultaneously as part of FAIRE. Each track is a live log: exercises done, experiments run, things built, resources that actually held up. The intent is twofold — compound understanding for me, and a structured starting point for anyone who wants to begin. If you want to get started on any of these, the what-to-build and where-to-start sections will tell you how I approached it from zero.
AI
Pre-training & next-token prediction · supervised fine-tuning · RLHF & RLAIF · instruction following · prompt engineering & chain-of-thought · tool use & function calling · agentic architectures & harness design · multi-agent systems · evaluation & benchmarking · mechanistic interpretability · alignment & safety · frontier model engineering
Generative Modeling
Autoregressive models · variational autoencoders · generative adversarial networks · normalizing flows · diffusion models & score matching · stochastic differential equations · flow matching & continuous normalizing flows · optimal transport · energy-based models · latent variable models · sampling & inference methods · generation quality & evaluation
Representation Learning
Self-supervised pretraining · contrastive learning (SimCLR, MoCo, CLIP) · masked autoencoding · joint embedding architectures · JEPA & predictive coding · latent space geometry & structure · disentangled representations · transfer & few-shot learning · multimodal alignment · vision-language models · audio & speech representations · probing & representation evaluation
Neural Networks & Deep Learning
Feedforward & convolutional networks · recurrent networks & LSTMs · attention & transformers · graph neural networks · optimization algorithms (SGD, Adam, variants) · regularization & generalization · training dynamics & loss landscapes · normalization (batch, layer, group) · scaling laws · neural architecture search · theoretical foundations of deep learning · efficient inference & deployment
Statistical & Probabilistic Machine Learning
Probability theory & measure · maximum likelihood estimation · Bayesian inference & prior design · conjugate models · MCMC & sampling methods · variational inference · directed & undirected graphical models · Gaussian processes · mixture models & EM algorithm · uncertainty quantification & calibration · probabilistic programming · approximate inference
Reinforcement Learning
Markov decision processes · dynamic programming & Bellman equations · Q-learning & DQN · policy gradient theorem · actor-critic methods · proximal policy optimization · model-based RL & world models · exploration strategies · offline RL & imitation learning · RLHF & reward modeling · multi-agent RL · game theory & equilibria
Attention, Memory, Reasoning and Continual Learning
Attention mechanisms (self, cross, sparse, linear) · transformer architectures · positional encodings & long-context methods · efficient attention variants · chain-of-thought & scratchpad reasoning · tool-augmented & multi-step reasoning · working memory & neural cache · retrieval-augmented generation · memory-augmented networks (Hopfield, NTM, DNC) · continual & lifelong learning · catastrophic forgetting & mitigation · task-incremental & class-incremental learning
Causal and Statistical Inference & Modeling
Structural causal models · potential outcomes framework · average & heterogeneous treatment effects · randomized experiments · observational studies & confounding · instrumental variables · difference-in-differences · regression discontinuity · mediation & moderation analysis · causal discovery (PC, GES, NOTEARS) · counterfactual reasoning · causal representation learning
Algorithms and Systems for AI
Algorithms & complexity analysis · data structures (trees, graphs, heaps, hash maps) · dynamic programming & search · graph algorithms · database systems & storage engines · vector databases & ANN indexes (FAISS, HNSW) · GPU & accelerator programming · data, model & pipeline parallelism · distributed training frameworks · mixed-precision & quantization · pruning & knowledge distillation · efficient architectures & operators · inference serving & batching · ML compilers (XLA, Triton) · data pipelines at scale · MLOps & experiment tracking
Complexity, Cognition and Natural Intelligence
Complex dynamical systems · nonlinear dynamics · emergence & self-organization · adaptive control · cognitive science & philosophy of mind · natural intelligence & neuroscience · embodied cognition · perception & attention · scientific discovery · applications in life, biological & physical sciences