Ongoing series
Each one is a single argument unfolded chapter by chapter.
Bailian model platform: prompt engineering, fine-tuning, agents, and evaluation.
Production-grade ML on Alibaba Cloud — DSW, DLC, EAS, Designer, QuickStart, end-to-end.
Infrastructure, networking, and the platforms ML actually runs on.
OS, networking, compilers — the substrate beneath everything.
Algorithms by pattern, with worked solutions.
The geometry and computation that underlies all of ML.
Practical Linux — shell, processes, networking, and performance.
Deriving the algorithms — no hand-waving.
Modern NLP — language models, embeddings, transformers, and beyond.
From classical ODE methods to neural ODEs.
PINNs, neural operators, and the math behind learned PDE solvers.
From classic CF to GNN, GCSAN, HCGR, and modern multi-objective ranking.
Foundations of RL: MDPs, policy gradients, actor-critic, and offline RL.
Building infrastructure-as-code agents: planning, validation, and apply loops.
Statistical and deep methods for forecasting at scale.
Domain adaptation, fine-tuning, and representation transfer.