$ cat course_outcomes_v1.log
pid:2026 · cohort:01 · April 2026
 ██████╗ ██████╗ ██╗   ██╗██████╗ ███████╗███████╗
██╔════╝██╔═══██╗██║   ██║██╔══██╗██╔════╝██╔════╝
██║     ██║   ██║██║   ██║██████╔╝███████╗█████╗
██║     ██║   ██║██║   ██║██╔══██╗╚════██║██╔══╝
╚██████╗╚██████╔╝╚██████╔╝██║  ██║███████║███████╗
 ╚═════╝ ╚═════╝  ╚═════╝ ╚═╝  ╚═╝╚══════╝╚══════╝
 ██████╗ ██╗   ██╗████████╗ ██████╗ ██████╗ ███╗   ███╗███████╗███████╗
██╔═══██╗██║   ██║╚══██╔══╝██╔════╝██╔═══██╗████╗ ████║██╔════╝██╔════╝
██║   ██║██║   ██║   ██║   ██║     ██║   ██║██╔████╔██║█████╗  ███████╗
██║   ██║██║   ██║   ██║   ██║     ██║   ██║██║╚██╔╝██║██╔══╝  ╚════██║
╚██████╔╝╚██████╔╝   ██║   ╚██████╗╚██████╔╝██║ ╚═╝ ██║███████╗███████║
 ╚═════╝  ╚═════╝    ╚═╝    ╚═════╝ ╚═════╝ ╚═╝     ╚═╝╚══════╝╚══════╝
$ labsoft-ai report --cohort=01 --format=detailed

Software Development
Processes Powered
by AI Agents

// An evaluation of an eleven-respondent cohort, conducted at the close of the inaugural delivery, April 2026.

Respondents
N = 11
Modules
06
Method
Anonymous
Recommendation
91%
would recommend the course to a colleague without reservation.
Instructor Rating
4.91/5
ten of eleven respondents marked the maximum.
AI in Pedagogy
4.82/5
approval of agentic instruction as a teaching paradigm.
Feedback <24h
82%
of respondents received instructor feedback inside one working day.
I — Mandate

Modern practice, taught the way the industry actually works.

// The course operationalises six engineering disciplines under a single, agent-driven workflow. Respondents were asked which of those disciplines they now feel they understand.

Test-Driven Development — RED · GREEN · REFACTOR
10 / 11 · 91%
Git Workflow — branching, commits, pull requests
9 / 11 · 82%
CI / CD — GitHub Actions, Docker
7 / 11 · 64%
Multi-agent Workflow
7 / 11 · 64%
User Stories & BDD
6 / 11 · 55%
Software Architecture — arc42, C4 model
3 / 11 · 27%
The most current technologies, expertise and dedication of the professor — it was an honour and a pleasure to attend such a course.
Cohort respondent · 25 April 2026
II — Reception

A unanimous endorsement of agentic instruction.

Introducing AI agents into the curriculum
Q4 — 1 (Poor idea) → 5 (Excellent idea)
4.82/5
Nine of eleven respondents rated the integration the maximum score.
0
1
0
2
0
3
2
4
9
5
Did Kiro CLI accelerate your learning?
Q5 — 1 (Not at all) → 5 (Significantly)
4.64/5
Eight reported significant acceleration; only one outlier at the median.
0
1
0
2
1
3
2
4
8
5
Utility of the automated AI reviewer
Q11 — 1 (Useless) → 5 (Highly useful)
4.55/5
Six awarded the maximum, validating the reviewer as core scaffolding.
0
1
0
2
1
3
3
4
6
5
Quality of course materials
Q13 — 1 (Poor) → 5 (Excellent)
4.73/5
Eight respondents awarded the maximum mark for README, slides and exercises.
0
1
0
2
0
3
3
4
8
5
III — The Value of the Agent

What working with agents actually changed.

Most-cited benefit · 11 of 11
100
11 of 11

Focus shifts from typing to understanding.

Every single respondent named this as a core benefit of the agentic workflow.

Faster Execution · 9 of 11
82
9 of 11

Tasks complete substantially faster.

The agent compresses cycle time from intent to working code.

Process Insight · 4 of 11
36
4 of 11

Context & harness engineering, in practice.

A meaningful minority cited deeper insight into agent configuration itself.

Configuration Insight · 4 of 11
36
4 of 11

Better grasp of process through agent configuration.

Configuring agents — not just prompting them — is itself a teaching mechanism.

An excellent topic — integration of AI, which is treated as a bogeyman in other courses but is, in industry, one of the principal aids. Compared to my current workplace, this was a faithful preview of the environment.
Cohort respondent · 25 April 2026
IV — Voices

In the respondents' own words.

// Open-text responses, presented verbatim. Translations from Serbian kept conservative.

Use of AI agents, minimal hand-typing of code, and the fact that the subject is highly applicable.
On what they liked most
The way the course was divided into modules, and the work with the agents.
On what they liked most
Newest technologies, the professor's expertise and dedication — it was an honour to attend. Many thanks to the professor.
On what they liked most
Working with AI tools that we will increasingly use, and the proper way of working with Git.
On what they liked most
Excellent topic. Integration of AI, which is treated as a bogeyman elsewhere but is in industry one of the principal aids. Compared to my current workplace, this was a faithful preview of the environment.
On what they liked most
The 'gamification' of the exercises, and the table on which we tracked our progress.
On what they liked most
It should appear in earlier years of study, because there is no other course that introduces students to GitHub.
On suggested changes
The only weakness, I believe, is the shortage of time, since most people are also working. Perhaps the course should be extended to four weeks.
On suggested changes
I would not change a thing — a very good course.
On suggested changes
Application of the entire learned process to a more complex task at the very end.
On suggested changes
// return 0;

Eleven cohort members. Six modules.
One conclusion: a curriculum that treats AI agents as colleagues,
not as novelties — and is graded accordingly.

// A report prepared from anonymous respondent data.