If you want to understand what an institution believes about talent, look at how it selects students.
The test design is never neutral. What you choose to measure is what you believe predicts success. What you leave out is what you believe can be taught, or isn’t worth selecting for at the door.
The KNET has three sections: Learnability and Comprehension, Problem Solving, and Communication. It’s 120 minutes, online, designed by senior tech executives for assessing a student’s fitment for a career in software. The test doesn’t follow a subject syllabus. No specific preparation is required beyond taking the mock test available in the portal.
Most families hear that and ask when the next slot is. The question I find more interesting: why these three sections, and not something else?
What most admissions tests are actually measuring
Subject-based entrance exams measure what students have already learned. Physics, chemistry, mathematics: content from secondary school that a well-prepared candidate has accumulated over years of study. The implicit theory is that students who’ve covered more ground, absorbed more content, and held up under pressure on a structured paper will do better in an engineering programme.
That’s not a wrong theory. For some purposes it works. For a software engineering programme with a particular design philosophy, it has a specific limitation.
In 1963, the psychologist Raymond Cattell published research that cognitive scientists and psychometricians have been building on since. He described two broad categories of intelligence, distinct but related.
Crystallized intelligence is the accumulated store: the concepts, formulas, and patterns you’ve learned and practiced until they’re automatic. It grows with education and experience, and subject-knowledge tests measure it fairly well.
Fluid intelligence is different. It’s the ability to reason with new problems. To extract structure from unfamiliar material. To figure out the pattern in something you’re encountering for the first time. It doesn’t depend heavily on what you’ve been taught before. Standard knowledge-based tests don’t measure it cleanly. Someone with a strong accumulated knowledge base can compensate for weaker fluid reasoning on a syllabus-based test. Someone with strong fluid reasoning but patchy subject coverage will tend to underperform their actual capacity.
Cattell’s distinction matters for software engineering because the field changes fast. Learning a new framework, reading documentation for a language you haven’t used before, debugging a system you didn’t build: these are fluid-intelligence tasks, not crystallized-knowledge tasks. A student who enters with high subject-test scores but low capacity for fluid reasoning hits a ceiling early. The KNET’s design reflects that judgement.
The three sections map onto that reasoning directly.
Learnability and Comprehension
This section asks whether you can encounter something new and extract its logic. Not whether you’ve already seen the concept. Whether you can work out what it means from what’s in front of you.
That’s a measure of fluid reasoning. It’s also a measure of what engineers call reading comprehension under real conditions: process unfamiliar material carefully, reason from the evidence in front of you, reach a conclusion. The ability doesn’t depend on your syllabus. It depends on how you think.
Kalvium looks for students who may not be top scorers but who show grit, self-awareness, and genuine capacity to grow. The Learnability section is the test’s attempt to find out whether that capacity is real rather than asserted.
Problem Solving
Unfamiliar problems, by design. The question isn’t “can you recall the method?” It’s “can you figure out the approach?”
Every working day, a software engineer encounters problems without pre-written solutions. They reason about the structure, identify the constraints, find the path through. You can’t practise your way to that using a fixed set of past papers. The Problem Solving section is trying to find out whether you can do it at all. Whether, when the problem is new and the answer isn’t recognizable, you stay in it or you don’t.
That capacity is what the section is measuring. Prior study can sharpen it. It can’t substitute for it.
Communication
This one surprises students sometimes. Why is verbal English in a software engineering admissions test?
Because software engineering isn’t a solo activity. Engineers write documentation, read error messages, explain design decisions to people who didn’t make them, and review code written by someone else under time pressure. The inability to communicate clearly isn’t a soft deficiency. It’s a ceiling on how effective an engineer can be, regardless of how well they reason in private.
The decision to include Communication in the test is a statement: we think the ability to express what you’ve found is part of what we’re selecting for. Not writing style. Clarity of thought expressed in language. A student who can think but can’t make another engineer understand what they’ve found will hit a professional ceiling the moment the work gets collaborative. Which is almost immediately.
What the full selection process is looking at
The KNET is one of three components. Admission to the Kalvium B.Tech CSE programme is based on merit and involves a Psychometric Assessment, the KNET, and an In-Person Interview.
The Psychometric Assessment comes first. It doesn’t test knowledge. It profiles how a student prefers to learn and work: how they respond when they’re stuck, whether they seek feedback or protect themselves from it, what their approach to ambiguity looks like. It adds a behavioural layer to the reasoning data from the KNET.
The In-Person Interview is where the character layer enters directly. Not a polished pitch. A direct conversation about who the student is and what they’ll do when things get hard.
Hunger shows up when someone talks about what they’ve been building without being asked. Grit shows up when they describe something that went wrong and didn’t stop there. Self-awareness shows up when they’re asked what they’re not good at and they answer honestly.
The students Kalvium is looking for are ready to show up with hunger, discipline, and curiosity. They may not be top scorers. But they show grit, self-awareness, and capacity to grow. The interview is the place in the process where those qualities can actually be seen.
The test tells you something about reasoning capacity. The interview is looking for something the test can’t measure.
What this design reveals
The selection process is a theory of engineering talent made operational.
Read together, the three KNET sections and the In-Person Interview say: we believe that fluid reasoning, communication capacity, and the character variables of hunger and grit predict engineering performance better than subject-recall scores alone.
That’s consistent with what the research on skill acquisition says. Anders Ericsson spent decades studying how people get good at hard things. His core finding: expertise is the product of deliberate practice, sustained engagement with challenge just beyond your current ability, with real feedback on what’s working. Given that, selecting for the capacity to engage in that kind of learning matters more than selecting for what someone already knows.
I want to be honest about where the evidence gets thin. The claim that the KNET measures fluid reasoning reliably is a design claim. It’s calibrated through iteration over admission cycles, not through independent psychometric validation. Whether the specific items within each section actually capture what they’re intended to capture is something the team refines over time. The three-section structure reflects a considered judgement. I hold the specific implementation with more uncertainty than the design philosophy itself.
A note on logistics
Registration for KNET is ₹1,200. It covers the full selection process: Psychometric Assessment, KNET, and In-Person Interview. Tuition is paid separately and directly to the partner university where the student enrols. Kalvium doesn’t collect tuition.
Nine partner universities for Admission Year 2026-27 offer the Kalvium B.Tech CSE programme on the basis of KNET selection. A single score applies across all nine. Slot dates aren’t published in advance. Phases open as seats are released.
Where to read more
If you want the patient, step-by-step explainer of what KNET is, how each component of the selection process works, and what families can expect after registration, the full guide is what is KNET: a complete guide for students and parents.
For the broader argument about what’s missing from most engineering programme designs, and the learning science reasons Kalvium’s structure is different, why most engineering programmes don’t produce engineers is where to start.
The test design is an admissions mechanism. It’s also a statement about what Kalvium thinks engineering capability actually is.
Arvind is Head of Programme Design and Delivery at Kalvium. He writes about the cognitive science of learning and how it shows up in the design of a live programme, grounded in named research and honest about where the evidence is still being worked out. Read more from Arvind or browse the B.Tech category.