DocuLens AI: What I learned building an AI reviewer in 24 hours
A field note on using AI-driven development to ship and deploy a source-grounded document review app without hiding the tradeoffs.
Read note @juancruzalbelo · Blog
Case studies & field notes
The systems behind
the portfolio.
Privacy-safe write-ups on production RAG, AI QA automation, LLM evaluation, agent workflows, and the engineering choices behind them.
Case studies decisions, architecture, boundaries
RAG retrieval, grounded prompts, refusals
Harness OpenSpec, subagents, verification
Archive
More notes
Mi Salud: a small app for keeping medical studies in one place
How I built a small private web app to organize personal medical studies, upload PDFs and images, and share temporary access during doctor visits.
Read →Ornith-1.0: the model trying to learn its own coding harness
An explanatory review of Ornith-1.0, DeepReinforce's open-source family for agentic coding based on self-scaffolding: what it is, why it matters, and where to stay cautious.
Read →Capytail: routing local de skills para agentes de IA
Capytail es una herramienta local para registrar skills, seleccionar capacidades relevantes y cargar solo el contexto que un agente necesita.
Read →Cómo implementamos una Advisor Strategy en Kilo Code
Una nota técnica sobre cómo adaptamos el patrón de Advisor de Claude Code en Kilo Code usando subagents read-only, benchmarks con GLM 5.2 Cloud y MiniMax M3, y una separación clara entre executor y advisor.
Read →Fine-tuning was not enough: how a local 14B model reached practical parity with a single-pass frontier baseline
A field note on a local LLM POC for Playwright test automation: why fine-tuning collapsed, what recovered it, and how retrieval, critique, repair, and verification closed the gap.
Read →Shipping an AI Portfolio with a Grounded Bot
How I shipped this portfolio as a small production AI product: Astro, MiniMax M3, curated retrieval, VPS deployment, and basic abuse hardening.
Read →I Built Two RAG Systems for DJs — Presskits and Set Analysis
A public, privacy-safe look at two production RAG systems I built inside PK Studio: one for creating artist presskits, and one for analyzing DJ sets and music-industry workflows.
Read →My AI Development Harness: Pi, OpenSpec, Subagents, and a Senior-Architect Workflow
A field note on how I configured my AI coding workflow around Pi, el Gentleman, Gentle AI, OpenSpec, subagents, model routing, skills, restore paths, and review gates.
Read →