Blog
Public technical articles on AI systems, evaluation, analytics, and production tradeoffs.
Writing
Public notes on system design, evaluation, and deployment tradeoffs.
The writing is the public companion to the project pages: why a system was built, what was measured, and which tradeoffs mattered once the work had to become usable.
Case-study style writing on LLM systems, forecasting, analytics, deployment choices, and engineering tradeoffs that held up once a prototype had to become usable.
Start with the four featured articles, then use the year-grouped archive to browse the rest without falling into one long scrolling wall.
Coverage
What the archive covers.
LLM and evaluation notes cover prompt optimization, retrieval, observability, and the difference between plausible outputs and defensible systems.
Applied ML writing covers finance, forecasting, and analytics systems where a model only matters if it changes a decision.
Product and deployment notes focus on what happened when a notebook had to become a public demo, dashboard, or production-minded workflow.
The archive is intentionally practical: more implementation and tradeoffs, less generic theory recap.
Featured reading
Start with four recent articles.
The top row is the fast entry point. The full archive sits below, grouped by year so it reads like an archive instead of one long feed.
Mar 24, 2026
Operational Workforce Risk Intelligence from Public Signals
Introduction: Public signals are only useful once they become incidents
Read article
Mar 24, 2026
Designing Creator AI as a Backend-First Platform
Introduction: generation quality is not the whole product
Read article
Mar 24, 2026
Building SlideBench for AI-Generated Slides
Introduction: generation tools need an evaluation layer of their own
Read article
Mar 24, 2026
Building Service-Oriented Document Intelligence
Introduction: Why this needed more than a single QA app
Read articleArchive
Browse the full archive.
16 articles across 2026 back to 2023, organized newest-first inside yearly groups.
