The semester schedule arrives the week before Year 2 begins. Most students scan it twice.
In Semester 1, the subject list had contained seven things, none of which were what they’d expected when they arrived. In Semester 2, seven more. Both semesters had a clear shape: the exercise had a right answer, the build had a brief, the DOJO problem had a correct solution. The work was hard, but it was bounded.
The Semester 3 list is different. Running alongside Database Management Systems, Object-Oriented Programming, and Introduction to Generative AI Systems is a subject listed simply as Internship / Project 1. That’s the structural change Semester 3 is built around. It’s the point where the brief stops being the boundary and starts being the starting point.
Why this semester marks the shift
For the first two semesters, every piece of work had parameters. A LiveBooks exercise with a defined input and a correct output. A DOJO problem where the right answer existed and the system confirmed you’d found it. A Capstone Challenge with a scope and a Friday deadline. Even when the problem was genuinely hard, it was solvable within a predictable range of outcomes.
Semester 3 is when that changes.
The Internship / Project 1 module doesn’t replace the academic subjects. It runs alongside them, applying the technical foundation from the first two semesters to work that doesn’t resolve into one correct answer. Database Management Systems arrives at the same time as real data modelling decisions. Object-Oriented Programming arrives at the same time as real systems that need to be designed and defended. That’s not a coincidence in the sequencing. It’s the reason the combination works the way it does.
The Semester 3 curriculum
The Software Product Engineering track in Semester 3 covers: Database Management Systems, Object-Oriented Programming, Introduction to Generative AI Systems, Values and Ethics in Technology, Financial Literacy, Environmental Sciences, and Internship / Project 1.
Students on the AI and Future Technologies track study the same core subjects but replace Financial Literacy with Mathematical Foundations for ML.
Introduction to Generative AI Systems is running in Semester 3 for the first time in a context where students are also building something with real constraints. That changes how the subject lands. Learning a technical domain while you’re applying it to a project isn’t the same as learning it as a standalone course. The project keeps asking a question the course has to answer: how does this actually fit into what I’m building?
Values and Ethics in Technology and Financial Literacy aren’t padding around the technical subjects. By Semester 3, students are making decisions that affect something beyond their own grade. An architectural choice about data modelling affects everyone who’ll use the system. A product direction decision has consequences further down. Those two subjects are building the context that makes the technical decisions harder to get wrong.
The four work tracks
From Semester 3 onward, students move into one of four tracks.
Simulated Work places students in in-classroom sprints structured to replicate a company engineering environment. There’s a brief, a team, and a delivery cadence. The mentor’s role shifts from someone who confirms whether an answer’s correct to someone who reviews whether the work’s production-ready. That’s a real change in the kind of feedback loop students are in. Students who’ve spent two semesters in a room where the question is “is this right?” are now in a room where the question is “does this hold up?”
Internship places students with tech-first companies. The work is real in the direct sense: it’s inside a real company, it affects real systems, and someone there cares whether it works.
Open Source Projects places students in global codebases, contributing pull requests. This track tests a specific capability that Year 1 is designed to build: can a student read a codebase they didn’t write, understand its architecture, identify where a contribution fits, and then defend the change in a pull request review conducted by people they’ve never met? The first two semesters build the technical foundation. The Open Source track is where that foundation faces its first genuinely external test.
Build AI-native products is the entrepreneurship track. Students here aren’t contributing to something that already exists. They’re building their own product, which means they’re also responsible for the product decisions, not just the execution. It’s designed for students who arrived with a specific idea and the drive to ship it, or who developed one during Year 1.
The Internship / Project module runs every semester from here: Semester 4, 5, 6, 7, and 8 each include an Internship / Project component that deepens and extends the work. Semester 3 isn’t the only exposure. It’s the first.
How the squad changes
The squad structure from Year 1 continues. What changes is what the squad is doing inside it.
In Semester 1 and 2, the most useful squad conversations had a structure: here’s where I’m stuck, here’s what I’ve tried, here’s where I think the problem is. The squad helped members reach the correct output. The problem was bounded, so the conversation had an endpoint.
By Semester 3, the squad is working through problems that don’t resolve that way. An architectural decision doesn’t have one right answer. A product direction question doesn’t either. The squad’s job shifts from helping each other reach the correct output to helping each other reason through tradeoffs. It’s a harder conversation to run well. It’s also the kind of conversation that matters.
This catches some students off-guard. The work in Year 1, however hard it felt, was structured to be resolvable. Semester 3 is the first semester where a student hits a genuine “it depends” and has to produce the answer themselves.
What DOJO looks like at this stage
DOJO continues from Semester 1 through all four years. The daily coding practice opens Semester 3 the same way it opened Semester 1. The belt progression runs in parallel: six belt levels per language across Java, C++, JavaScript, and Python, each tested, not assumed.
By Semester 3, the belt progression has been running for over a year. What’s different isn’t the format. It’s the relationship a student has with the session.
Daily practice in Semester 1 was building the habit. By Semester 3, the habit’s there. The DOJO session is now about maintaining precision under real project pressure: keeping the coding practice running while a different kind of cognitive load from the work track sits alongside it.
Students who let daily practice slide when the project work picks up tend to find the gap later. The technical round in a placement conversation asks for something specific. If it hasn’t been touched in months, the gap shows.
What HEROS tracks at this stage
HEROS, Kalvium’s real-time learning system, runs across all four years. Mentors have a dashboard showing where each student is moving forward, where they’re stalling, and where a gap is opening before the student reports it themselves.
In Year 1, most gaps were predictable. A student who struggled with recursion. A student who couldn’t read a database schema independently. The curriculum produced consistent failure patterns, and mentors learned to recognise them early.
By Semester 3, the gaps are harder to predict, because the work is harder to predict. A student might be technically strong but can’t scope a project brief without expanding it into something undeliverable. A student might build well on their own but isn’t able to communicate what they’re building clearly to the rest of the team. These aren’t syntax gaps. They’re judgment gaps. HEROS surfaces the signal; the mentor follows it before it becomes something the student has to raise themselves.
What the end of Semester 3 produces
By the end of Semester 3, a student’s completed the first Internship / Project module.
Not a course exercise. Not a practice build. Something that ran under real constraints, and that they can describe in specific terms when someone asks: what have you built?
That’s the question Semester 3 is designed to give a student an answer to. The semesters that follow build on it. Each subsequent Internship / Project module adds to the portfolio. By Year 4, the question “walk me through what you’ve done” doesn’t feel like a test. It’s a description of something that started in Semester 3 and kept going.
For the learning science behind why building real work from Year 2 changes what students retain and can do, the research on how engineers actually learn makes the case directly. For the full picture of the programme from KNET through to the final year, the complete Kalvium guide for families has everything a family needs before deciding.