API Design in Node.js vs Web Performance Fundamentals, v2

The current version (v5) is a roughly 10-hour, end-to-end build of a production REST API: Express routing and middleware, a Postgres database with migrations, JWT-based authentication and authorisation, TypeScript throughout, runtime schema validation with Zod, error handling and integration testing with Vitest, finishing with a deploy to Render. Reviewers repeatedly describe it as "comprehensive" and as covering "all the important backend topics" in a single coherent project. The one structural criticism, raised by a workshop attendee, is that the database-choice discussion (SQL vs NoSQL) arrives later than it should, and a few exercises bleed code meant for later steps into earlier ones.

Scott Moss — a senior engineer at Netflix and a two-time Y Combinator founder — is the most consistently praised element across our entire sample. Learners describe him as explaining "each and every concept and line of code in an easy-to-understand and easy-to-follow way," and one blogger notes his "super relaxed, but brilliant mad scientist vibe that makes learning feel comfortable." A reviewer of his related Node.js course calls his teaching "engaging and informative, making complex topics accessible to learners of all levels." No reviewer in our sample criticises his clarity; the only instruction-adjacent note is occasional ambiguity about where an exercise is meant to stop.

The course is not standalone-purchasable: it is included in a Frontend Masters subscription (monthly or annual), which also unlocks the entire catalogue including Scott Moss's other Node, Next.js and AI courses. Reviewers who already subscribe treat this course as one of the highest-value backend titles on the platform; one blogger who tried 20+ backend courses lists it among his top recommendations. The subscription model means it is excellent value for active learners but poor value for someone who wants only this one ~10-hour course and nothing else — there is no one-time purchase option.

There is no graded feedback, peer review or instructor marking — this is a recorded workshop, not a cohort course. What learners get instead is a well-structured GitHub repository with per-lesson branches and exercise solutions, which several reviewers single out as excellent for "quick lookups" and for checking their work. In-person workshop attendees got live Q&A, but on-demand viewers do not. The exercise-scope ambiguity noted by one reviewer ("it was often a little unclear where we were supposed to stop") is the main friction point in the self-check loop.

This is the course's strongest dimension. The stack it teaches — Express, Postgres, JWT, TypeScript, Zod, Vitest, deploy to Render — maps directly onto what working backend teams actually ship in 2026, and one reviewer explicitly notes the API design patterns "apply to Java, Python, Go, Node.js and other backend technologies," not just Node. Multiple learners report feeling "more confident about building APIs" and "what I'm doing in Node.js and TypeScript" immediately afterward. The production-deployment ending is the part reviewers most often credit for closing the gap between tutorial code and shippable code.

The course covers the full stack of modern web performance knowledge in approximately six hours — a tight, well-curated curriculum that avoids the padding common in longer Udemy-style recordings. It opens with the psychology of perceived performance (drawing on David Maister's "Psychology of Waiting Lines"), then moves through Core Web Vitals metrics (LCP, CLS, and INP, which replaced FID as an official Google signal in March 2024), measurement APIs, synthetic testing with Lighthouse and WebPageTest, Real User Monitoring with CrUX, and finally concrete optimization tactics for each metric — TTFB, FCP, LCP, layout shift prevention, and interaction latency. The October 2024 update brought the curriculum fully current with the INP transition, so learners are not working with obsolete tooling or metrics. What sets the content apart from generic performance tutorials is its insistence on real user data before optimization. The recurring lesson — "focus on your worst metric, fix the most basic thing first, and confirm with RUM" — is a workflow, not just a collection of tips. The course also links performance work directly to business outcomes: conversion rate, SEO rankings, and user retention statistics are woven into the justification throughout, which gives engineers the language they need to advocate for performance work with non-technical stakeholders. The GitHub workshop repository (113 stars, 87 forks as of mid-2026) demonstrates the exercises have genuine uptake in the developer community.

Todd Gardner is one of the most credentialed practitioners teaching web performance in any format. As co-founder of TrackJS and Request Metrics, he has spent years building commercial RUM tools and working directly with thousands of development teams on performance problems — a background that produces very different teaching than a course built purely from documentation. His Frontend Masters blog articles (published November 2024 on INP and February 2025 on image optimization) extend the same practical, measurement-first methodology into the broader developer community. Student feedback collected from the official course page is unusually consistent in citing his teaching clarity as a differentiator. Anurag Bhandari wrote "Wonderfully planned and executed. Such a heavy topic explained in such simple terms." Pedro Antônio Pereira called it "a masterclass in how to understand web performance." Ryan Davidson, recommending it broadly, wrote: "Great blend of breadth and depth in the performance space. All web engineers — backend or frontend, junior or senior — should be taking this course!" The breadth of that recommendation — extending to backend engineers — reflects how Gardner anchors the course in concepts (HTTP caching, server response times, rendering pipelines) that apply regardless of whether you write CSS or database queries.

The course is included in a Frontend Masters subscription, priced at approximately $39/month or $390/year (annual plans have carried a 17-20% discount in 2025-2026). A seven-day free trial is available. For a subscriber who accesses even two or three courses, the per-course cost is minimal, and the Web Performance Fundamentals course is among the most immediately actionable in the entire library — the techniques taught apply to any existing project without needing to rebuild or refactor an architecture. Deeptiman Mallick's testimonial captures the evergreen value: "This course is like a bookmark to come back to when we're working on performance." Unlike courses that you complete once and set aside, the performance metrics and tools taught here — Lighthouse, WebPageTest, CrUX — are ones practitioners return to on every new project or optimization sprint. The value proposition is strongest for working developers with a real codebase to optimize; the course is less compelling as a purely theoretical exercise for learners with no project to apply it to. There is no standalone purchase option, so non-subscribers must commit to at least one month of the full subscription.

The hands-on component uses a real Node.js + Express e-commerce project — "Developer Stickers Online" — which is deployed to multiple regions and available on a global CDN with HTTP/3 support, making it possible to test real network conditions rather than synthetic localhost scenarios. Students work with actual Lighthouse scores, WebPageTest waterfalls, and CrUX field data rather than simulated metrics, which bridges the gap between tutorial and professional practice. The workshop's focus on a single, realistic project (rather than a series of disconnected toy exercises) is appropriate for the subject: web performance work lives in real measurement data, and the course correctly models that. The project repository's 87 forks suggest learners actively run the exercises rather than just watching. The main limitation is scope: the project is a static storefront, so learners working primarily on Single Page Applications, server-rendered frameworks like Next.js, or complex state-heavy UIs will need to extrapolate some of the tactics. There is no separate project tier with more complex application types.

Web performance is one of the highest-ROI skills a working web developer can acquire in 2024-2026: Core Web Vitals are a confirmed Google Search ranking signal, and the business case studies cited throughout the course (eCommerce and SaaS conversion improvements correlated with performance gains) are drawn from real production data. Gardner's background running RUM tools for thousands of teams means the tactics are not academic — they are the same ones practitioners reach for when a client's LCP score drops. Multiple student reviews specifically noted immediate applicability: Álex Castelo wrote that the course made them realize "how easy it can be to boost a website's performance exponentially," and Yuganshu Mohan distilled the practical lesson as "focus on the worst metric and perform the most basic fixes." The methodological frame — measure with real user data first, then fix, then confirm — transfers directly to professional performance audits. Taran Bains, writing an extended reference post from the course material in August 2025, used it as the foundation for a comprehensive web performance guide, which is a reliable signal that the content holds up as a reference long after the initial watch-through.