Thoughts, tutorials, and insights
JEV from TypeSafe AI opens a new category: a System One model that returns typed decisions with probabilities and calibrated confidence, in a tenth of a second and at a fraction of LLM cost. What that means for decision trees, routing, lead scoring and guardrails, and how the architecture differs from a regular language model.
Read More →
Everyone says "train the AI on your data". RAG does nothing of the sort: no weights change, your data stays in your database, and answers are grounded at question time. What actually happens, why it is better for business, and when fine-tuning does make sense.
Read More →
AI FOMO makes companies buy platforms nobody uses. The adoption path that works is boring: map processes, pilot with measurable ROI, RAG on company knowledge, evals and guardrails, team enablement. What it saves, and how I help companies get there in weeks.
Read More →
CRM per seat, a leads tool, a forms tool, and Google Sheets holding it together. The economics of building your own tools changed: weeks instead of quarters, AI in the middle instead of a paid add-on, and data you actually own. Where to build, where to keep buying, and the math.
Read More →
A practical guide to WebMCP: what it is, how to enable it in Chrome, how LLMs communicate with it, best practices, and realistic time savings for product and engineering teams.
Read More →
A comprehensive, battle-tested guide to choosing between MongoDB and PostgreSQL: scalability, performance, transactions, indexes, managed services, and real-world use cases for SaaS and eCommerce. Includes decision frameworks, diagrams, and code examples.
Read More →
Agent skills make AI assistants reliable: repeatable workflows, safe defaults, and less prompt churn. Here’s how Claude Code skills and Vercel’s React rules help teams ship faster with fewer regressions.
Read More →
A practical playbook for improving day-to-day delivery: connect Slack, Jira, Monday.com, and GitHub using MCP-based AI agents for notifications, triage, status sync, code review feedback loops, and automated follow-ups.
Read More →
A behind-the-scenes breakdown of this portfolio: Next.js 16.1.3 App Router, next-intl (he default + /en), Sanity CMS, Clerk auth for AI chat, Resend email, and full SEO with sitemap/robots/JSON-LD, optimized for static builds.
Read More →
How I ship AI chat features safely: Clerk-gated access, OpenAI ChatKit sessions, prompt/response guardrails, and performance-minded client loading in a Next.js 16 App Router codebase.
Read More →
Patterns I use to keep a Next.js App Router project maintainable: strict TypeScript, CMS-driven content with Sanity, stable GROQ queries, and practical trade-offs when the CMS evolves faster than the code.
Read More →
How I approach real-time UX (chat, live updates): WebSockets, auth/session boundaries, backpressure, and how to keep performance predictable in modern Next.js + Node.js systems.
Read More →
A practical checklist for speeding up Postgres-backed apps: query plans, indexes, caching with Redis, and what to watch for when you run on serverless Postgres like Neon.
Read More →
What I learned moving into leadership: aligning product + engineering, setting quality bars (CI/CD), mentoring, and shipping reliably while still staying hands-on technically.
Read More →