Papers, blog posts, and downloads.
Open knowledge from the Sariro team. Read the research, grab the templates, steal the checklists. No email gate on most of it — we believe AI education should be free.
The library
8 resources ·updated bi-weekly
A practical framework for evaluating LLM applications in production
Why most "LLM evals" are theater, and a 5-layer framework for measuring what actually matters: correctness, latency, cost, safety, and user satisfaction.
12 prompting patterns every AI engineer should know by heart
Stop vibes-prompting. These 12 patterns — from chain-of-thought to self-consistency — will make your prompts reproducible, testable, and shippable.
The Sariro RAG pre-launch checklist (PDF, 2 pages)
A printable 2-page checklist covering the 24 things you must verify before shipping a retrieval-augmented generation system to real users.
Teaching thinking, not typing: a pedagogy for AI literacy
A research-grounded argument for why AI education must prioritize systems thinking over syntax — with data from 3 cohorts and 1,200 students.
Vector DB shootout: pgvector vs Pinecone vs Weaviate in 2026
We benchmarked the three most popular vector databases on the same RAG workload. The winner depends on three things — and they're not what you think.
Sariro AI Builder Starter Kit (GitHub template)
A ready-to-clone Next.js + Python starter with RAG, evals, and a deploy button. Used by every Sariro cohort as their day-one scaffold.
On the reliability of autonomous AI agents: a field study
We deployed 7 agent architectures across 4 production workloads for 90 days. Here's what broke, what held, and what we'd never ship again.
How to design a CSTA-aligned AI curriculum for grades 9–12
A walkthrough of the Sariro schools curriculum — module by module, learning objective by learning objective — with a free downloadable syllabus.
Three pillars. No fluff.
Applied research, practical engineering, open tools. That's the entire editorial scope — promise.
Applied research
Field-tested findings from real Sariro cohorts and production deployments. We publish what works — and what didn't.
Engineering blog
Practical walkthroughs, patterns, and benchmarks. Read it Monday, ship it Tuesday.
Open tools
Templates, checklists, and starter kits under permissive licenses. Take them, fork them, use them at work.
One email. Every other Tuesday.
The latest research, the best new tools, and one practical thing you can ship this week. No spam, no sales, no nonsense.