Technical writing samples

OpenAPI Spec & Live Response Verification Script

A Python script that validates live API responses against OpenAPI schemas before documentation is written — preventing drift, hallucination, and outdated code examples.

Read the article

Roomly developer guide: build on the student housing API

A practical guide to building with Roomly's proposed student housing API, from your first request to a full dashboard experience.

Read the article

Roomly API specification: student housing data model

A proposed Roomly API for student housing data, listing availability, locations, and marketplace information for developers and integrators.

Read the article

Roomly caretaker guide: activate, list, and withdraw

Activate a caretaker account, list properties, manage reviews, and withdraw earnings on Roomly without getting stuck in the flow.

Read the article

Roomly student guide: sign up, book, and post roommate ads

A straightforward guide to the main Roomly student flows: signing up, booking housing, and posting or browsing roommate ads.

Read the article

5 Sub-3B Small Language Models You Can Run on Your Laptop CPU (No GPU Required)

You don't need a cluster GPU to run local AI. Compare 5 sub-3B Small Language Models (SLMs) that run on consumer laptop CPUs with 1–4GB RAM at 4-bit quantization.

Read the article

Why Do Autonomous AI Agents Fail? The Architecture Bottlenecks

AI agents don't fail because models are dumb. They fail because architectures let them get confused—context drift, infinite loops, and tool hallucinations. Here's how to fix each.

Read the article

How Kubernetes HPA scales

Kubernetes autoscaling looks simple, but the HPA loop, formulas, and default settings are where the real behavior is decided.

Read the article

LLM production cost: APIs vs self-hosted GPUs (2026)

There isn't one honest number for LLM cost. It depends on traffic, latency, model choice, and how much infrastructure you want to own.

Read the article

Serverless vs containers for AI workloads: when to choose each (2026)

The real question is not which is better, but which workload shape you actually have. A practical framework for AI deployment decisions.

Read the article

How RAG works: query, retrieve, and generate

A mechanics-first explanation of retrieval-augmented generation: how the query, retrieval, and generation steps work together in practice.

Read the article

Build a RAG pipeline with Node.js and TypeScript

A practical step-by-step tutorial for building a retrieval-augmented generation pipeline in Node.js and TypeScript, then hardening it with real error handling.

Read the article

Recursive zk-SNARKs: how Mina keeps its blockchain succinct

Recursion lets Mina prove an entire chain with a compact proof. Here's how the mechanism works and why it matters.

Read the article

How Mina keeps a blockchain under 22KB

Mina compresses a blockchain into a tiny proof. Here's what that means in practice and why the 22KB number matters.

Read the article