Assess your cohort against industry hiring demand.
Score a cohort on the tasks industry hires for, then give every student a plan and the time to act on it.
Portfolios built on real work
Problems from industry partners
Evidence per student and cohort
cadra · sample report
cadra × YOUR UNIVERSITYGRADED
A. Sample Student
B.TECH CS · BUILD THE AI
Problem
61%
Dev
38%
Overall
46%
Roadmap
PER STUDENT · WHAT TO FIX, IN ORDER
AI use in build24%
Eval design38%
Solution architecture67%
Cohort · 87 students
BASELINE · ONE PASS
Community presence16%
AI use in design35%
Movement · T1 → T2
RE-BASELINE · END OF TERMILLUSTRATIVE
46%→63%
+17 PTS ACROSS COHORT
Illustration: a sample Cadra report cycling through a student capability profile, a per-student roadmap, an illustrative cohort baseline, and illustrative term-over-term movement.
AI capability is what employers screen for, and no transcript records it.
Employers want to see how a student works, measured early enough to change it.
Claims outrun proof
Students list AI tools on a CV. Employers have no way to tell who works well with them.
Coursework grades the output
Assignments are marked on what was produced. Industry cares how the work got done, and with what judgment.
The gap surfaces too late
Most students discover what they are missing during placement season, with no runway left to fix it.
EU AI Act, Article 4: organisations must ensure AI literacy in their workforce. Graduates who can demonstrate it are easier to hire.
Which makes it a measurement question before it is a training one.
Two signals, read against what industry demands.
Proof of work
What they have already built, in public.
+
Live build
What they build now, watched end to end.
vs
Hiring demand
The bar industry sets, and the rubric Cadra scores against.
CADRA SETS THIS
SCORING AGENT READS THE TRAIL · ALL THREE AGENTS RUN THE BUILD
The two tracks read different surfaces, because the work is different.
BUILD THE AI
Engineering cohorts
CS · DATA SCIENCE · ENGINEERING
Can they build and ship an AI system that holds up?
Build a working prototype from a one-line product request, then justify the scope cut to ship it.
Measured on
Product judgmentAI tool masteryStakeholder reasoningSpeed to prototype
One cohort is judged on what they can build, the other on how well they think with AI in the loop. Each gets its own task and its own rubric.
It runs across a semester, with time built in to act on what it finds.
The same shape for both tracks. Only the task and the rubric change.
BASELINE
Where the cohort starts
Both signals are read: the public trail of what they have built, plus a live task built for their track.
RECORDED AS THEY WORK
ONE WEEK, NO SCHEDULING BURDEN
→
ROADMAP + BRIDGE
What to fix, and the help to fix it
Each student gets their own gaps in priority order. The bridge programme then runs for the cohort, built from the gaps the batch shares and sized to your academic calendar.
ROADMAP PER STUDENT
BRIDGE PER COHORT
→
RE-BASELINE
What changed
The same task, the same rubric, at the end of term. The movement between the two runs is the proof.
BEFORE AND AFTER
EVERY STEP AIMED AT WHAT INDUSTRY HIRES FOR
Baseline against the industry rubric
Industry use casesIndustry mentors
Industry-ready portfolio
↺Re-baseline feeds the next cohort's baseline.
Evidence at two levels: per student, and across the cohort.
The same evidence, read two ways.
PER STUDENT
Individual capability profile
Scored against the rubric industry hires on, and backed by the work itself.
A roadmap of what to fix, in what order, with enough runway before placement season to fix it.
Capability that stands up in an interview, built on tasks modelled on the real job.
PER COHORT
Department gap report
Where the cohort stands by track, early enough in the year to act on it.
Which gaps are systemic and worth teaching, and which are individual and worth coaching.
Re-baselining across a term produces a before and after, so progress is visible.
Sample reportsOne student and one cohort, anonymized.
Select a track
PER STUDENT
Sample Student
B.TECH CS · BUILD THE AI
27%OVERALL
Problem solving53%
Profile29%
Development9%
COACHING · WHAT TO DO NEXT
Build a consistent project pipeline, and engage with the AI community
Commit weekly for three months, small fixes included, to build a visible track record.
Publish two more AI projects with READMEs, evaluation metrics and deployment notes.
Contribute to an open-source project, or publish two Kaggle notebooks with the analysis written up.
Community presence and AI use in design have the most ground to make up. The bridge course is built from this ranking, worst first, so the batch spends its term on what it most needs.
No MBA cohort has been assessed yet. These figures show the shape a first run would produce, on the rubric in the section above.
PER STUDENT
Sample Student
MBA · WORK AI-NATIVELYILLUSTRATIVE
49%OVERALL
Product judgment64%
Stakeholder reasoning57%
AI tool mastery41%
Speed to prototype33%
ROADMAP · WHAT TO FIX, IN ORDER
PER COHORT
MBA cohort
BASELINE · SEMESTER 1ILLUSTRATIVE
44COHORT MEAN / 100
Speed to prototype33%
AI tool mastery41%
Stakeholder reasoning57%
Product judgment64%
The systemic gap sits in building rather than framing. This cohort can reason about the decision well before it can produce anything to test it against.
One engineering college, one pass, measured end to end.
87STUDENTS BASELINED IN ONE PASS, ONE WEEK
6INTERNSHIPS CONVERTED FROM THE TOP TEN
All 87RECEIVED A CAPABILITY REPORT WITH COACHING RECOMMENDATIONS
An AI/ML branch at an engineering college in Hyderabad, assessed in one pass: a live simulation plus a review of each student's GitHub, Kaggle and LeetCode record. Every student received an individual capability report with coaching recommendations, and the department received the cohort gap report.
COHORT THROUGHPUT, ONE PASS · INTERNS NOW AT PARTNER STARTUPS
Chaitanya Bharathi Institute of TechnologyHYDERABAD
Let's talk about your cohort.
A short conversation about which programmes to start with, what the scenarios would look like, and what the cohort report would tell you.
