Resume keywords for software engineers, with real job posts

Resume keywords for software engineers in 7 roles, checked against real job posts, and why a skill shown in a bullet counts more than one listed under Skills.

By the fastcv team · 28 October 2026 · 9 min read

The resume keywords that matter for a software engineer are the tools, practices and job title that the specific job post names, spelled the way the post spells them, and shown in a bullet where you used them. A keyword sitting only in your skills line is weaker: fastcv's job-match checker gives a must-have skill full credit when it appears in a bullet, role or project, and only three-quarters credit when it's merely listed. It also reads "Postgres" as PostgreSQL and "k8s" as Kubernetes, which a recruiter's literal search may not.

The checker's weights are public: must-have skills 45%, nice-to-have skills 10%, other repeated key terms 20%, years of experience 15% and job title 10%, as described on the free resume and job description checker. Below are keyword groups for seven roles, checked against real posts on company careers pages, plus the rules that decide whether a keyword helps you.

What counts as a keyword on a software engineer resume

A keyword is a word or short phrase a recruiter or screening tool can search for: a language (Go), a framework (Spring Boot), a datastore (PostgreSQL), a practice (CI/CD, code review) or a title (Backend Engineer). It counts only in the form someone types into a search box.

That matters because recruiters search, not just read. Naukri's recruiter database, Resdex, has separate "Any", "All" and "Exclude" keyword fields and Boolean AND, OR and NOT, so a recruiter can ask for resumes containing "Kafka" AND "Java" and NOT "intern". If your resume says "message queues", it never appears in that list.

Not every keyword sets you apart. In the Stack Overflow 2025 Developer Survey, 66% of respondents had done extensive work in JavaScript over the past year, 58.6% in SQL and 57.9% in Python, and 71.1% had used Docker, so those are table stakes. The keywords that move you up a shortlist are the specific ones a post repeats: Kafka, Spark, Jetpack Compose, Terraform.

Software engineer resume keywords by role

Backend posts ask for a language plus datastores and queues; frontend posts for TypeScript, accessibility and performance; data posts for SQL, Spark and an orchestrator. The example bullets below are made up; use your own numbers.

Backend engineer

Keywords: Go or Java, Node.js, REST APIs, microservices, PostgreSQL or MySQL, Redis, Kafka or RabbitMQ, AWS, Docker, system design.
From a real post: Paytm's Backend Senior Software Engineer post asks for "2 to 5 years of server-side application development experience with Golang", then lists "Java/Node", "Kafka/RabbitMQ" and "Redis/Aerospike".
Example bullet: "Moved order events from cron jobs to Kafka consumers in Go, cutting status-update delay from 5 minutes to under 2 seconds."
Full sample: the backend engineer resume example.

Frontend engineer

Keywords: React or Vue, TypeScript, JavaScript, CSS, accessibility, web performance, Core Web Vitals, testing (Jest, Playwright), design systems.
From a real post: GitLab's Senior Frontend Engineer post makes architecture, browser behaviour, testing, accessibility and performance the requirements, and calls Vue, React, Angular and TypeScript "Helpful, but not required".
Example bullet: "Rebuilt the checkout form in React with keyboard and screen-reader support, cutting form errors from 900 to 300 a week."
Full sample: the frontend engineer resume example.

Full-stack developer

Keywords: React, Next.js, Node.js or Spring Boot, TypeScript, REST APIs, GraphQL, PostgreSQL or MongoDB, authentication (OAuth), CI/CD, Docker.
What posts reward: one stack in depth, the whole path in one bullet.
Example bullet: "Shipped team invites end to end: a Node.js API, a PostgreSQL schema and a React screen, used by 140 teams in the first month."
Full samples: the full-stack developer resume example and, for MongoDB-first posts, the MERN stack developer example.

Data engineer

Keywords: SQL, Python or Scala, Spark, Kafka, Airflow, dbt, a warehouse (Snowflake, BigQuery, Redshift), data modelling, data quality, streaming.
From a real post: Meesho's SDE III Data post asks for "Java, Scala, or Python, and expertise in SQL", then Kafka, Spark, Hive, Delta Lake, Presto/Trino and Airflow, and offers "Airflow, Dagster, or Prefect" as alternatives.
Example bullet: "Replaced nightly CSV loads with a Spark Structured Streaming job into Delta Lake, bringing dashboard freshness from 24 hours to 15 minutes."
Full sample: the data engineer resume example.

DevOps engineer or SRE

Keywords: Kubernetes, Docker, Terraform, AWS or GCP, CI/CD (GitHub Actions, Jenkins), Helm, Prometheus, Grafana, Linux, incident management, observability.
What posts reward: uptime, deploy frequency, time to recover and cloud cost.
Example bullet: "Moved 40 services to Kubernetes (EKS) with Helm and Terraform, cutting deploy time from 25 minutes to 6."
Full sample: the DevOps engineer resume example.

Machine learning engineer

Keywords: Python, PyTorch or TensorFlow, scikit-learn, LLMs, NLP or computer vision, feature engineering, MLflow, model deployment, Docker, a cloud (AWS, GCP or Azure), SQL.
What posts reward: a model in production, the metric it moved and how it was served.
Example bullet: "Fine-tuned a PyTorch ticket classifier and served it behind FastAPI, routing 60,000 support tickets a month without manual triage."
Full sample: the machine learning engineer resume example.

Android developer

Keywords: Kotlin, Java, Jetpack Compose, MVVM, Coroutines, Room, Hilt, Retrofit, Firebase, Gradle, unit and UI testing, Play Store.
What posts reward: shipped apps with installs, ratings, crash rates or startup time.
Example bullet: "Rewrote the onboarding flow in Jetpack Compose with Kotlin Coroutines, cutting cold start from 2.1 s to 1.2 s."
Full sample: the Android developer resume example.

Shown versus listed: how fastcv's checker counts a keyword

fastcv's checker scores work you show (role titles, bullets, projects, added sections such as achievements) apart from what you only list (headline, summary, skills, education). The table compares five ways a keyword can appear.

On your resume	What fastcv's checker does	What to do
"Kafka" only in the Skills line	Counts the must-have at 75% credit	Add a bullet that shows what you built with it
"Kafka" inside an experience or project bullet	Full credit	Keep it, ideally with a number
"Postgres" when the post says "PostgreSQL"	Counts it (Postgres and psql are aliases)	Write "PostgreSQL" anyway; a recruiter's Boolean search may be literal
"k8s" or "EKS" when the post says "Kubernetes"	Counts it as Kubernetes	Write "Kubernetes (EKS)" so both forms appear
Post says "Go or Java" and you have only Java	Doesn't mark Go as missing	Don't add Go to fill the gap

The Skills line is an index, not proof. It still earns three-quarters of the credit, but the bullet is what a recruiter reads after the search finds you.

Use the job's exact spelling

fastcv's checker knows 256 skills and tools with their common aliases, so "golang" counts as Go and "GKE" counts as Kubernetes. Recruiter search may not be that forgiving: Resdex's "All" field returns resumes containing every keyword typed, so a search for "Golang" is looking for that word.

Posts are also inconsistent. Paytm writes "Golang"; GitLab writes "representational state transfer" where most posts write REST. Copy the post's spelling where it's true for you, and add the common form once in brackets: "Go (Golang)", "Kubernetes (EKS)", "PostgreSQL".

The job title is a keyword too: it's 10% of fastcv's score, which reads "developer" as "engineer" and "SDE" as "software engineer" and ignores seniority words like "Senior" or "II". If your official title was "Member of Technical Staff", your headline can say "Backend engineer" as long as that's the work you did. The full method for rewording a resume to one post is in how to tailor your resume to a job description.

Never list a keyword you can't defend

Every keyword on your resume is an interview question you've agreed to answer. If "Kubernetes" is on the page, expect "how would you debug a pod stuck in CrashLoopBackOff?" Run three tests before a keyword goes on:

Used it, not watched it: you wrote, deployed or debugged something with it, at work, an internship or a project.
One bullet of proof: you can point to a bullet that shows it. If you can't write one, it's not ready.
Ten minutes of questions: you could talk about it for ten minutes, including one thing that went wrong.

fastcv's AI tailoring applies the same line: requirements your resume can't back are listed as "can't claim" instead of being written in. If tailoring a resume to a post shows a gap, the honest fix is a project, not a keyword.

Keywords for freshers and campus placements

A fresher's keyword list is shorter and different. Stripe's Software Engineer, New Grad post for Bangalore says "We work mostly in Java, Ruby, JavaScript, Scala, and Go", then adds that new languages "can be learned if the fundamentals and general knowledge are present". The language list describes their stack; it isn't a filter you must pass.

For on-campus and off-campus drives in India, the keywords that recur are fundamentals: data structures, algorithms, OOP, DBMS, operating systems, computer networks, SQL, Git, plus one or two languages you actually code in. Show them through projects, coursework and coding-profile results, as in the software engineer fresher resume example. The resume format for freshers covers where each section goes.

When a keyword list is the wrong tool

When the post isn't keyword-led. GitLab's frontend post above makes frameworks optional. Adding "Angular" to match a list would answer a question the post never asked; its real keywords are accessibility, performance and testing, shown in bullets.

When a checker misreads the post. fastcv's checker marks a line nice-to-have only when it uses a phrase like "nice to have", "preferred", "a bonus", "is a plus" or "ideally", so GitLab's "Helpful, but not required" line under "What you'll bring" counts as must-haves. It also reads "Java, Scala, or Python" as Scala-or-Python, so a Python-only resume still shows Java missing; read the post yourself before chasing every red item.

When nobody searches. A referral, a founder reading your GitHub, or a research role judged on papers won't run a keyword search. There, one strong project link matters more.

When you need the process, not the list. This page is a reference. If you have one post open and want to rework your resume for it, use the tailoring guide instead.

Frequently asked questions

Do keywords in the skills section count?

Yes, but less. In fastcv's checker a must-have skill that's only listed earns three-quarters of the credit of one shown in a bullet, and a recruiter who finds you through a search still reads the bullets to check it.

How many keywords should a software engineer resume have?

There's no fixed number. Cover the post's must-haves that are true for you, each once in a bullet and once in Skills, and stop; repeating a keyword five times adds nothing for a reader.

Should I write "K8s" or "Kubernetes"?

Write the form the post uses, and add the other in brackets if it's common: "Kubernetes (EKS)". fastcv's checker reads both, but a recruiter's literal search may not.

Do soft skills count as resume keywords?

Rarely as search terms. Show them through what you did ("led the migration with three teams", "wrote the on-call runbook") rather than listing "team player" or "communication".

Can I add keywords for tools I've only done a course in?

Only if you built something with it. Otherwise put the course under Certifications and keep the tool off your Skills line until a project shows it.

Should my Naukri and LinkedIn profiles use the same keywords?

Yes. Recruiters search profiles by keyword, so the skills and titles on your profiles should match your resume and the roles you're applying for.