Anonymised at the institution's request
A 620-student engineering department in western Maharashtra
The problem. The training and placement cell had bought three years of generic
“aptitude and coding” training. Students cleared the online rounds in reasonable numbers and then
collapsed in technical interviews — the post-mortem sheets kept naming the same four topics:
complexity analysis, indexing, process versus thread, and anything involving a network.
What we did. A single sixteen-week cohort across all six courses, sequenced
DSA → OOP → DBMS → OS → CN → System Design, with weekly targets set by the department and progress
gated module by module. Fortnightly live sessions were run by the instructor who wrote each module.
Two mock interview drives were held, six weeks and two weeks before the season, with a written
feedback sheet per student.
What the department measured. Module completion was the only metric we asked them
to hold us to, because it is the only one we control. By week sixteen, 71% of enrolled students had
completed all six courses and 88% had completed the first four. The cell's own interview post-mortems
stopped listing indexing and process-versus-thread as recurring failure points.
What we will not claim. Placement percentages moved that year. We are not going to
tell you we caused it — the same season saw the cell change its recruiter mix and add a
communication programme. Anyone who attributes a hiring outcome to one training vendor is selling you
something.