Eyeing the ‘Future of Work,’ AWS Expands Kiro for Students Program
AWS seeks to encourage structured thinking and practical training by getting its agentic development tool in front of more students
Using Kiro by AWS, developers can build and ship real engineering work using AI agents. In March 2026, the company began a pilot program, offering students at a few universities free access to Kiro. The pilot was a success, and AWS is expanding it to include students at 132 universities in 18 countries around the world. Each student will have 1,000 Kiro credits per month, no strings attached.
Expanding Kiro for Students
The original program was in just two countries at 11 universities, and thousands of students used it to create. “We see a lot of students use these tools to build something that’s truly changing lives,” says Darko Mesaros, distinguished developer advocate, AWS. “We wanted to give other students access to these tools to be able to work in their fields and expand their skills and be ready for the future of software development—really the future of work.”
The universities included in the expanded program were those with students and faculty who requested the Kiro team bring the program to their campus, who were logistically ready to take on the program, and have an eager build community. The company expects to continue adding institutions over time.
Impacts and Expectations for Kiro for Students
During a time when enrollment in computer science programs is declining, the workforce is undergoing a massive change, causing some computer and information science professionals to question their future. A program like Kiro for Students can help people learn to apply what they know in new ways.
“The skills people learn when it comes to software architecture, systems, and scaling, and resiliency—all of those things still apply. They just happen to generate code with an agent instead of typing it out themselves,” Mesaros says. Change has been consistent throughout the history of technology, and agentic tools are another change—but foundational knowledge remains important.
“Coding was and is a part of software development,” he says, “But coding is not software development.” Understanding infrastructure, reliability, scale, the customer base, the total addressable market and all of those kinds of factors is still crucial, and having a tool to help with coding doesn’t change the need for that knowledge.
Mesaros thinks one of the most important advantages of Kiro for Students is that it will help students focus their efforts and approach projects in a more structured and rigorous manner.
“It’s super easy to just build stuff without structure, but spec-driven development forces you to sit down, think about what you want to build, read the whole requirements document and the design document and the list of tasks your agent will execute, and then actually agree and confirm that yes, this is exactly what I want to build,” Mesaros says. By approaching projects in a more intentional and structured way, students can build more useful and less “sloppy” software.
Training and Student Builder Groups
Mesaros recently traveled and met with student groups across Canada and Brazil and participated in and observed workshops where students were learning spec-driven development and best practices for building with AI, including with Kiro and agentic tools more generally. He says the student builder community around AWS is particularly strong and the content those community groups tend to seek out is related to Kiro.
“Kiro has wonderful documentation, and an excellent blog with a lot of practical examples,” Mesaros says, “and Kiro Crew was actually built with Kiro.” Kiro Crew provides a real-world, practical example of a production-grade, massively used product. Other opportunities for community and learning include a Discord group, conferences, meetups and workshops for students to learn and be involved in the community.
Students have access to four surfaces through the program:
1. The IDE, where agents write the code, run tests and verify correctness while the students review and make decisions.
2. The CLI, which includes custom agents for complex code bases and allows students to automate workflows.
3. Kiro Web, for those times when Kiro can’t be downloaded.
4. Kiro Crew, a persistent, open-source space that remembers context, learns as students go and coordinates tools and workflows.
Success Stories From the Pilot Program
Students aren’t only using Kiro to build demonstration projects as portfolio samples. The projects they are completing are, as Mesaros says, “truly changing things.” For example, one group created an accessibility app that turns visual content on the web into audio content for visually impaired people.
Another group of students built a voice-activated system called ChanseyCare that connects patients in their homes to medical professionals. “Our system creates an earlier layer of remote triage where a clinician can check in, assess the patient through video and audio, and escalate [to EMS and 911] only when needed,” the students say.
One of the exciting elements that has emerged from the pilot program is that students from a diverse range of disciplines want to learn how to use Kiro. Many computer science majors participate, but students from adjacent disciplines like data science are also interested in learning how to use agentic tools. “Because, yes, it is a software development tool,” Mesaros explains, “but it can also do a lot of other things.”
As an example, Mesaros describes a tool he built for himself to keep up with social media, which is a key part of his job. He says it scours the internet and builds a report that lets him monitor mentions and follow up on conversations. “That’s not software engineering, but it is incredibly useful for my work,” he notes, adding that people who learn to use Kiro are applying it for reasons beyond software engineering, as he has.
It’s Still About Attention
As he contemplates the possibilities that AI unlocks for software developers, Mesaros thinks back to the 2017 Google research paper, “Attention Is All You Need,” which introduced the transformer architecture that kicked off the current wave of generative AI.
“Weirdly enough, almost 10 years later, for us to be successful as software developers with agentic tools, attention is all we need,” Mesaros says.