Writing

Notes from systems and research work.

Explanations, comparisons, and operating lessons. Older writing stays visible, but evidence-heavy case studies remain primary portfolio record.

01

Technical notes.

Foundational pieces from earlier portfolio, migrated into current design system.

01
January 25, 2026 · Natural language processing
Understanding Transformer Models in NLP
Why attention replaced recurrence, how transformer blocks work, and where their advantages and costs matter.
Read note ↗
02
January 20, 2026 · Machine learning
Building AI Models with Limited Data
A practical sequence for transfer learning, augmentation, regularization, and evaluation when labels are scarce.
Read note ↗
03
January 30, 2026 · Deep learning
LSTM vs GRU: Choosing Right RNN Architecture
Architecture differences, runtime tradeoffs, and practical experiment that should decide between gated recurrent models.
Read note ↗