Oğuzhan Kurum

I recommend GitHub's workflow guide when you want to understand how build, test, and deployment gates live beside the code they protect.

Railway’s public networking guide is a practical reference for custom domains, TLS, and production service exposure.

I recommend Railway's official CLI guide when you want repeatable deploy, logs, variables, and service operations without leaving the terminal.

The official fundamentals guide explains how Cloudflare handles domains, DNS records, traffic, certificates, caching, analytics, and account security. It is useful when an incident crosses the boundary between application code and the network edge.

The official introduction shows how protocol buffers, generated clients, streaming, deadlines, and language interoperability fit together. It is a clear starting point for deciding when typed RPC is a better boundary than ad hoc JSON endpoints.

Start with the official overview when observability tools feel like unrelated dashboards. OpenTelemetry explains the vendor-neutral model for signals, context propagation, collectors, instrumentation, and exporting telemetry across services.

The official concepts guide connects workloads, services, storage, configuration, security, policies, scheduling, and cluster administration. It is the right starting point before copying manifests whose runtime behavior you cannot predict.

The official DevTools documentation is worth browsing beyond Elements and Console. Network traces, performance recordings, memory tools, accessibility inspection, coverage, sensors, and protocol tooling can replace a surprising amount of guesswork.

Google publishes its Site Reliability Engineering books online. They provide a shared vocabulary for service levels, error budgets, monitoring, incident response, toil, capacity, and operating distributed systems without relying on folklore.

The official Stripe CLI guide covers installation, authentication, API requests, webhook forwarding, event triggering, logs, and fixtures. It is the fastest way to make a payment integration observable before real money is involved.

Use this official guide to understand build context, layers, cache behavior, multi-stage builds, BuildKit, secrets, and reproducibility. It helps turn a Dockerfile from a copied recipe into an artifact you can reason about.

Redis becomes much more useful once you think beyond key-value caching. The official docs connect strings, hashes, sets, streams, transactions, persistence, clustering, and client behavior to real design choices.

This official introduction is a good way to rebuild the fundamentals beneath Express, Nest, and build tools: the runtime, modules, asynchronous I/O, the event loop, package management, and common development patterns.

The overview explains why fetching, caching, synchronization, invalidation, retries, and background refresh belong together. Read it before building another collection of custom loading flags and duplicated request effects.

Zod is useful wherever static types meet untrusted runtime data. Its docs make schemas, parsing, refinements, transformations, error handling, JSON Schema, and ecosystem integrations easy to discover without hiding the underlying contract.

Use Pro Git when memorized commands are no longer enough. It builds a durable model of commits, branches, remotes, rebasing, internals, hooks, and collaboration, and the complete book is available online.

Read this short methodology when a service begins accumulating environment drift, hidden dependencies, stateful assumptions, or inconsistent deploy behavior. The language is old enough to be stable and concise enough to review in one sitting.

This course explains loading, rendering, images, fonts, JavaScript, resource hints, and Core Web Vitals as one performance system. It is a practical path from a slow-page symptom to the browser behavior causing it.

The WAI tutorials connect accessibility requirements to concrete page structures: images, tables, forms, menus, carousels, and page regions. They are far more useful during implementation than treating accessibility as a final audit.

This is the security reference to consult while designing authentication, sessions, passwords, uploads, APIs, SSRF defenses, logging, and secrets handling. Each guide is focused enough to apply during implementation and review.

Use the official Actions documentation for workflows, events, expressions, permissions, environments, reusable jobs, caching, and security. It is especially valuable when a copied workflow almost works but its trust boundaries are unclear.

The Playwright introduction gets you from installation to resilient browser tests quickly, then opens into fixtures, traces, screenshots, multiple browsers, mobile emulation, and CI. It is the reference to keep beside any serious web QA loop.

pnpm is worth learning beyond the install command. Its official docs cover workspaces, catalogs, overrides, patches, deployment, and strict dependency resolution without treating monorepos as an afterthought.

Use the Turborepo docs to get task dependencies, caching, filtering, environment variables, and CI behavior right. The payoff is not just faster builds; it is a monorepo whose execution model the whole team can reason about.

The Tailwind documentation is excellent for translating a visual decision into exact layout, spacing, typography, responsive, and state utilities. Keep it open while building instead of guessing class names from memory.

Drizzle is a strong fit when you want TypeScript inference without giving up the shape of SQL. Its overview explains the design, schema model, query APIs, migrations, and why the library stays close to the database.

When a query, transaction, lock, index, JSON operator, or extension behaves unexpectedly, go to the current PostgreSQL manual. It is precise, searchable, and unusually good at explaining both the feature and its operational consequences.

NestJS is easiest to understand through its official docs: modules, providers, controllers, guards, pipes, testing, and deployment are presented as one architecture rather than disconnected recipes.

Use the official Next.js documentation before relying on a blog post that may describe an older router. The guides connect routing, rendering, caching, data fetching, deployment, and upgrades to the current framework behavior.

The official React quick start is the cleanest reset for components, JSX, state, events, lists, and data flow. It is compact enough for a first pass and accurate enough to revisit when old React habits get in the way.

Start here when TypeScript stops feeling like annotated JavaScript. The official handbook explains everyday types, narrowing, generics, modules, compiler behavior, and the mental model behind the type system in one coherent path.

Fable 5 is one of Anthropic’s most well guarded models yet

Heykeline beton yetmez

Every app breaks. The difference is whether you find out from an angry customer or from a clean error report with stack trace and context already attached. Sentry is one of those tools you don’t appreciate until it saves you at 2 a.m.

Postgres, auth, storage, realtime, edge functions — the boring infrastructure you need to ship fast, wrapped in a developer experience that actually feels modern. If Firebase ever felt too closed or too magical, Supabase is the clean, open-source alternative worth building on.

If you rely on the mouse for everything, Raycast shows you a different way to use your computer. Launch apps, find files, run commands, use extensions, and move around with a few keystrokes. Once you get used to it, reaching for the mouse feels slow.

Node, npm, bundlers, test runners—Bun turns the whole JavaScript toolchain into one fast, unified tool. Install dependencies, run scripts, bundle code, and write tests with a CLI that feels instant. The first time you switch a project over and everything speeds up, you can't unsee it.

Whiteboard that feels like paper: zero login friction, hand-drawn style, instant share links. My default for architecture sketches.

Cut through the noise: the official, plain-English explanation of what Ethereum actually is and why it matters, straight from the source instead of a guy trying to sell you a coin. No price predictions, no "get rich" pitch, just a clear look at the technology. Whether you're curious or skeptical, start with the real thing.

No royalties, no license fees, no strings, just a real game engine you fully own. Godot has quietly become the go-to for indie developers who don't want to hand a cut of their game to a big company. Lightweight, friendly to beginners, and improving fast. If you've ever wanted to make a game, download this tonight.

Before Stripe, accepting money on the internet meant weeks of paperwork and gnarly APIs. Now it's a few lines of code and you're live. Their docs are the gold standard every other developer tool aspires to. Whether you're charging for a SaaS, a marketplace, or a side project, this is where you start.

This 2017 paper introduced the Transformer, the architecture behind ChatGPT, Claude, and basically every modern AI system. It's remarkably short and readable for something this consequential. Even if you skim it, understanding where "attention" comes from demystifies a huge amount of the AI hype. The primary source everyone cites but few actually read.

Quanta explains the frontier of math, physics, and computer science in a way that's rigorous but genuinely readable. No dumbing-down, no clickbait, just careful, gorgeous writing about the ideas reshaping how we understand the world. I save their long reads for the weekend.

This is where operators go for advice that actually helps. In-depth, no-fluff pieces on hiring, product, fundraising, and management, drawn from founders and leaders who have actually done it, not pundits guessing. Every article is long-form and genuinely useful. If you're building or running a company, it's the closest thing to sitting down with someone ten steps ahead of you.

Founder or not, these essays will change how you think about work, ambition, and building things people want. "Do Things That Don't Scale" and "How to Do Great Work" alone are worth more than most business books. Plain text, no paywall, no fluff, one of the sharpest minds in tech thinking out loud. Read one on a slow afternoon.

Design patterns used to be gatekept behind a dense, decades-old book. This site explains every one with clear diagrams, real code in multiple languages, and plain language. When you hit a gnarly bit of architecture and think "someone must have solved this shape before," this is where you find out how.

Most AI courses bury you in equations before you ever train a model. fast.ai flips it: you build working models from lesson one, then learn the theory once you care. Taught by Jeremy Howard, it's free and has launched a lot of careers. The single best on-ramp to actually doing deep learning.

Unlike tutorials that spoon-feed you, The Odin Project makes you build real things and figure out the hard parts, which is exactly how you actually learn. It's the closest thing to a free, self-paced bootcamp, and it's beloved for a reason. If you're serious about becoming a full-stack developer, this is the path.

Thousands of people have gone from zero to a developer job using freeCodeCamp, and it costs nothing. Thousands of free hands-on tutorials, full courses, and certifications that teach real fundamentals instead of chasing the framework of the month. If someone asks you how to start coding, send them here first.

The terminal barely changed for 40 years. Warp rethought it: blocks instead of an endless scroll, real text editing, AI that writes the command you can't remember, and shareable workflows. It sounds unnecessary until you use it for a day and going back feels ancient. If you live in a shell, treat yourself.

After years of heavy, sluggish project tools, Linear feels like someone finally respected your time. Keyboard-driven, instant, and opinionated in all the right ways. Small teams that switch to it stop dreading their tracker. If Jira makes you sigh, spend a week here.

It replaced Spotlight on day one and then slowly ate half my other apps. Launch anything, run clipboard history, manage windows, trigger scripts, all from one keystroke. The extension store turns it into whatever you need. If you're on a Mac and live on the keyboard, this is a genuine daily-life upgrade.

Auth, database, storage, and realtime, all on top of real Postgres you actually own, not a proprietary black box. I've shipped whole products where Supabase was the entire backend. It's the open-source answer to Firebase, and because it's just Postgres underneath, you never hit a wall you can't code your way out of.

Push to git, get a live URL. That's genuinely it. Vercel made deploying front-end apps so painless it changed how a whole generation of developers ships. Preview deployments for every pull request, an edge network out of the box, zero server config. If you're still SSHing into a box to deploy, let this free you.

Every front-end developer eventually realizes MDN is the source of truth. Not a random blog, MDN. Accurate, exhaustive, and kept current by the people who help build web standards. Bookmark it, and when you're stuck on how a CSS property or JS method really works, this is where you go.

Andrew Ng taught a huge share of the world's ML engineers, and this is where his courses and short, practical guides live. The free short courses on prompting, RAG, and building with LLMs are the fastest way to go from "I use ChatGPT" to "I build with these models." Clear, current, and refreshingly hype-free.

This is the tool that made running large language models on your own computer genuinely easy. One command and you've got Llama, Mistral, or dozens of other open models running locally: private, offline, and free. No API keys, no per-token bills, no sending your data to someone else's server. If you want to build with AI or just experiment without a cloud account, start here.

It's VS Code, but the AI isn't bolted on, it's the whole point. Highlight code and describe the change in plain English and it edits in place. The codebase-wide chat answers questions about your own project, not generic Stack Overflow answers. If you've outgrown copy-pasting from a chat window, this is the upgrade.

Whether you want to run a model locally, fine-tune one, or just see what's state-of-the-art this week, everything lives here. Hundreds of thousands of open models, datasets, and demos you can try in the browser before writing a line of code. I check the trending page every morning: it's the fastest way to know what's actually moving in AI, not just what's loud on social media.

I've tried every AI coding tool and this is the one that stuck. It runs right in your terminal, reads your whole repo, and edits real files instead of spitting out snippets you copy-paste. The difference is it understands context across your codebase, so it's less "autocomplete" and more like a junior engineer who never gets tired. If you write code for a living, spend one afternoon with it and you won't go back.