Skills index · Q3 2026 edition
The Most Requested IT Skills, Q3 2026
A ranked list of the skills employers actually asked for in the jobs they posted on iitjobs this quarter: what rose, what slipped, how the AI cluster is moving, what the US and India ask for differently, and which skills travel together. Updated every quarter from the live jobs data.
Skills Index · 8 min read
How to read this page
Every job posted on iitjobs carries a skills field that the employer fills in. This page counts those fields for every new posting between July 1 and September 25, 2026, and compares them with April to June. Reposts and refreshed listings are excluded, so a job is counted once, when it is first posted. Spelling variants are merged (React, ReactJS and React.js are one skill; AWS, EC2 and Lambda are one skill). A skill's share is the percentage of postings that list it. The average posting lists seven skills, so shares add up to far more than 100.
More than 18,000 new postings qualified this quarter. Two thirds are in the United States, about a quarter in India, and the rest elsewhere.
Python leads by seven points and has for two quarters. The interesting movement is everywhere below it.
The top 25
| # | Skill | Q3 share | Q2 share | Change |
|---|---|---|---|---|
| 1 | Python | 27.6% | 27.3% | +0.3 |
| 2 | AWS | 20.3% | 20.1% | +0.2 |
| 3 | REST APIs | 20.3% | 17.3% | +3.0 |
| 4 | SQL | 18.9% | 17.5% | +1.4 |
| 5 | Java | 17.9% | 18.0% | −0.1 |
| 6 | CI/CD | 17.2% | 15.7% | +1.5 |
| 7 | Azure | 11.7% | 12.3% | −0.6 |
| 8 | Agile / Scrum | 11.6% | 10.9% | +0.7 |
| 9 | JavaScript | 11.0% | 11.0% | 0.0 |
| 10 | Kubernetes | 9.2% | 8.6% | +0.6 |
| 11 | Git | 8.5% | 8.6% | −0.1 |
| 12 | GCP | 8.2% | 9.0% | −0.8 |
| 13 | React | 8.0% | 7.8% | +0.2 |
| 14 | Spring Boot | 7.9% | 7.6% | +0.3 |
| 15 | Microservices | 7.4% | 5.7% | +1.7 |
| 16 | Docker | 7.4% | 7.3% | +0.1 |
| 17 | .NET | 6.3% | 6.3% | 0.0 |
| 18 | HTML / CSS | 6.3% | 6.8% | −0.5 |
| 19 | Terraform | 6.2% | 5.3% | +0.9 |
| 20 | TypeScript | 6.2% | 5.4% | +0.8 |
| 21 | Jira | 6.1% | 5.5% | +0.6 |
| 22 | C# | 5.5% | 5.4% | +0.1 |
| 23 | Angular | 5.5% | 5.7% | −0.2 |
| 24 | Kafka | 5.1% | 4.8% | +0.3 |
| 25 | Generative AI / LLMs | 5.1% | 4.0% | +1.1 |
Ranks 26 to 50, in order: Azure DevOps, Jenkins, Databricks, ETL, SQL Server, Oracle, DevOps, PostgreSQL, Linux / Unix, Snowflake, AI / ML, ServiceNow, Power BI, Bash / Shell, Node.js, Networking (TCP/IP, BGP, DNS), PowerShell, JSON, Spark, PySpark, Data Modeling, GitHub Actions, Selenium, SAP, MySQL. Each of these appears in between 2.6% and 5.1% of postings.
What moved
The PyTorch and TensorFlow decline next to the RAG and LLM rise is the clearest signal in the data.
Employers are asking for fewer people who train models and more people who build on top of them.
Risers.
REST APIs gained three full points, the biggest move on the list, and Microservices gained 1.7. Read together, they say the same thing: employers are hiring people to build and connect services, not monoliths. CI/CD (+1.5) and Terraform (+0.9) rose with them, along with GitHub Actions and TypeScript. SQL gained 1.4 points, which is worth noticing in a quarter where every conversation is about AI.
The AI risers.
Generative AI / LLMs rose from 4.0% to 5.1% and entered the top 25 for the first time. RAG doubled, from 1.2% to 2.4%. Agentic AI nearly doubled, from 0.8% to 1.5%. GitHub Copilot doubled, from 0.6% to 1.2%. Databricks gained 1.1 points and Snowflake 0.7 on the data side.
Fallers.
GCP slipped from 9.0% to 8.2% and Azure from 12.3% to 11.7%, while AWS held. PyTorch (1.5% to 1.1%) and TensorFlow (1.4% to 1.0%) both fell in the same quarter that LLM and RAG rose. HTML / CSS, jQuery, MongoDB and Power BI eased slightly.
The AI cluster
Group every AI-related skill together (generative AI, LLMs, RAG, agentic AI, LangChain, MLOps, PyTorch, TensorFlow, model APIs, vector databases, Copilot and classic AI / ML) and 11.5% of this quarter's postings name at least one of them, up from 10.0% in Q2. The generative slice alone went from 5.3% to 6.8%.
Month by month, the line only goes one way.
| Month | Any AI skill | Generative AI | RAG | Agentic AI |
|---|---|---|---|---|
| April | 9.3% | 5.3% | 1.3% | 0.7% |
| May | 9.6% | 5.4% | 1.2% | 0.7% |
| June | 11.4% | 6.1% | 1.2% | 1.1% |
| July | 11.2% | 6.2% | 1.7% | 1.1% |
| August | 12.3% | 7.7% | 2.6% | 1.6% |
| September | 11.1% | 7.4% | 3.0% | 1.7% |
Inside a generative AI posting, the rest of the stack is predictable: Python appears in 62% of them, RAG in 29%, AWS in 27%, agentic AI in 18%. "AI Engineer" is now the seventh most common job title on the platform, ahead of .NET Developer and Python Developer. MCP, the model context protocol, appeared in postings for the first time this year and is still under half a percent.
Cloud: the requisition is a menu
27.5% of postings name at least one cloud. AWS is on 20.3% of all postings, Azure on 11.7%, GCP on 8.2%. The number that matters for candidates is the overlap: one in three cloud postings lists two or more providers. AWS and Azure together appear on 7.3% of every posting on the platform, and among postings that ask for AWS, 36% also ask for Azure and 25% for GCP.
Employers are not choosing a cloud in the skills field. They are listing the ones they run, and expecting the hire to move between them.
Data engineering is the largest single role
Data Engineer is the most common job title this quarter, ahead of Software Engineer and Java Developer. The stack around it is consistent. In postings that name data engineering as a skill, Python appears in 56%, SQL in 51%, AWS in 33%, ETL in 33%, Databricks in 32%, Snowflake in 24% and PySpark in 21%. Databricks was the fastest-rising data tool of the quarter.
The Java stack, and the third of it that also wants Python
Java is flat at 17.9%. Inside Java postings, Spring Boot is named 36% of the time, REST APIs 33%, Microservices 23%, and Python 32%. Nearly a third of Java requisitions now also ask for Python, which is a larger share than any framework except Spring Boot. .NET holds at 6.3% with C# at 5.5%.
United States and India ask for almost the same list
The top ten is nearly identical in both countries. Python leads both (27.5% in the US, 28.9% in India), followed by AWS, SQL, REST APIs and Java. The differences are in process and platform. Agile / Scrum is on 12.8% of US postings and 7.1% of Indian ones. CI/CD is 17.6% in the US against 14.1% in India. Azure runs slightly higher in India (13.2% against 11.2%).
Where the remote jobs are
15.7% of new postings are flagged remote, 19% hybrid and 44% on-site, with the rest unspecified. Among the top skills, Terraform (17.0% remote) and GCP (16.9%) are the most remote-friendly. Spring Boot (9.6%) and Angular (10.5%) are the least. Generative AI postings sit right at the platform average.
Skills that travel together
| Pair | Share of all postings |
|---|---|
| AWS + Python | 8.9% |
| Python + SQL | 8.2% |
| AWS + Azure | 7.3% |
| CI/CD + Python | 6.6% |
| Java + Spring Boot | 6.5% |
| Java + REST APIs | 5.9% |
| Java + Python | 5.7% |
| AWS + CI/CD | 5.4% |
| Docker + Kubernetes | 5.2% |
| AWS + GCP | 5.1% |
Among postings that ask for Python, AWS is named in 32%, SQL in 30% and CI/CD in 24%. Python is not a language requirement on this platform so much as the connective tissue of every other stack.
The ten most common titles
Data Engineer, Software Engineer, Java Developer, Business Analyst, DevOps Engineer, Full Stack Developer, AI Engineer, Java Full Stack Developer, .NET Developer, Python Developer. Seniority prefixes (Senior, Lead, Principal) are folded into the base title.
By category
| Category | Q3 | Q2 |
|---|---|---|
| Programming languages | 51.5% | 51.1% |
| Web and application frameworks | 37.1% | 35.0% |
| DevOps and platform | 32.9% | 30.9% |
| Data engineering and analytics | 32.4% | 31.0% |
| Cloud | 27.5% | 27.1% |
| Enterprise platforms (Salesforce, SAP, ServiceNow, Oracle) | 16.3% | 15.4% |
| Networking and infrastructure | 12.1% | 11.7% |
| Generative AI and ML | 11.5% | 10.0% |
| Testing and QA | 5.9% | 5.2% |
| Security and identity | 4.9% | 4.0% |
| Mainframe and legacy | 3.4% | 3.5% |
Every category except mainframe grew or held. DevOps, frameworks and AI grew the most.
What to do with this
If you are looking.
Python plus one cloud plus SQL covers the three largest overlaps on the platform. If you already have that, RAG and agentic AI are the fastest-growing additions and are still rare enough on resumes to stand out. If you are a Java developer, listing Python is no longer optional; a third of Java postings ask for it.
If you are hiring.
This list is only as good as the skills field. Postings that list the actual stack (the cloud, the framework, the data tools) match better in Search Talent and rank better in job search than postings that list adjectives. Seven skills is the platform average; the strongest postings list the ones a candidate would search for.
Next edition.
The Q4 2026 edition will be published in early January 2027 from the same method, with this page kept as the archive.
Search jobs by skill
Every skill on this list is a live search on iitjobs. Open the jobs board, type the skill, and see this quarter's postings for it.
Search jobs by skill