How do you read a skills list with fifteen technologies on it?

How do you read a skills list with fifteen technologies on it?

You sort it. The list arrives as a wall of names, and the shelves turn the wall into three or four groups: which names are languages, which are frameworks, which are tools, which are platforms. Once the names are grouped, the list starts answering questions. Where does this person live in the application? Is there depth anywhere? Does the list agree with the work history printed below it?

Fifteen names is normal, and it is normal for honest resumes as well as padded ones. An engineer with eight years of work has genuinely touched fifteen technologies, because every job adds a few. The skill in reading the list is separating the shape of a real career from the shape of a filter strategy, and the sort does most of that work for you.

Why is an unsorted list unreadable?

Because the names are different kinds of thing, and unlike kinds resist comparison. A raw list puts Python next to React next to Docker next to AWS as if they were four entries in one category. They are a language, a framework, a tool, and a platform, and each one implies a different amount and a different type of experience. Reading them left to right is like reading "French, Excel, negotiation, Salesforce" as one skill: every word is true and the sentence says nothing.

The unsorted list also hides the one thing you most want to see, which is concentration. Depth in a resume shows up as several related names clustered around one layer of the application. In an alphabetical or arbitrary ordering, that cluster is scattered across the line, and a genuinely deep backend developer looks identical to someone who did a tutorial on each of the same names.

How do you sort the wall?

Two passes. First by kind: mark each name as a language, a framework or library, a tool, a platform, or a practice, the categories from the shelves. Second by layer: place each one where it lives in the application, frontend, backend, database, or the infrastructure underneath. Fifteen names collapse into a handful of groups, and unfamiliar names can wait in a residue pile without blocking the read. Two passes on a resume takes about a minute once the shelves are familiar.

What the sort produces is a picture with a shape. Most groups will be thin, one or two names, and one group will usually be thick. The thick group is the candidate's home layer. A frontend cluster reading React, TypeScript, and Next.js is one developer with one specialty described three ways, because those names stack on each other. The residue pile is informative too: names you cannot place are usually niche tools from a specific job, and they tend to resolve the moment you read the work history.

What does a coherent stack look like?

A coherent list reads as a stack with some history attached: roughly one main pick per layer, plus a deep cluster at the home layer, plus a few older names that mark previous jobs. TypeScript, React, Node.js, PostgreSQL, and AWS is a complete tech stack for a web product, one choice per shelf, and it describes a full-stack developer. Add Redux, Next.js, and CSS depth to that list and the lean is visible: this person lives on the frontend half.

Coherence also means the list agrees with the jobs. A resume claiming deep Kubernetes should show a role where someone ran software in production at a scale that needs it. The strongest confirmation is a technology appearing twice, once in the list and once inside a bullet describing what was shipped with it. The list is the claim. The history is the evidence. When both name the same technology, the claim is very likely real, and when a listed name never appears in any job, it usually marks a course or a side project.

What is keyword stuffing, and why does it happen?

Keyword stuffing is listing technologies primarily so an automated filter matches the resume against a job description, rather than to describe experience. Developers know their resumes are screened by keyword before a human reads them, they compare notes about it publicly, and many respond by listing every technology they have ever touched. Given how the filters work, that response is rational, and it is common enough that a stuffed list says almost nothing negative about the person behind it.

What it does do is change what the list can tell you. A stuffed list has a recognizable shape after the sort: names spread across every layer, several competing picks on the same shelf, and no cluster anywhere. Four frontend frameworks side by side is the tell, because one deep framework plus a fast ramp is how real careers work, and four at once is how keyword matching works.

What does T-shaped mean?

T-shaped describes the skill profile of most experienced engineers: broad working familiarity across many areas, the horizontal bar of the T, with real depth in one, the vertical bar. Years of work generate the horizontal bar automatically. Every project touches neighboring layers, so a backend specialist of eight years has necessarily brushed against frontend code, databases, and deployment. Depth in one layer plus contact with the rest is what a career produces.

This is why a fifteen-name list can be completely honest. The horizontal bar alone can account for ten of the names. The reading question is never whether the bar is wide, since it always is. The question is whether the sort finds a vertical bar, one layer where the names cluster and the work history goes deep. A list with a clear vertical bar is a T-shaped engineer described accurately. A list that is all horizontal bar describes exposure, which is real and useful, and which prices differently than expertise.

Where does a GenAI experience line go?

Onto one side of the line drawn in building AI versus using AI, the same sort as every other claim on the page. A line reading "GenAI experience", "LLM integration", or "built AI features" almost always sits on the using side: the developer called a model through an API and built product around the answers. That is genuine, current, in-demand backend work, and the pool of people who can do it is large and growing.

The building side uses different words. "Trained models", "fine-tuning", "inference optimization", and named model architectures signal work on the model itself, and the pool holding those skills honestly is a small fraction of the pool holding the first kind. The two claims often wear the same "GenAI" label on a skills wall, and they describe different candidates at different prices. Sorted, the line is one of the most informative on the modern resume, because it is new enough that the work history behind it is short and specific, and a single project bullet usually settles which side the candidate is on.