Candidate placeability prediction
WorkstreamProduct thinking
A career team has limited time and needs a better starting point for where to focus it. The placeability model was a way to make candidate readiness more explicit rather than rely only on individual judgement. Before a data scientist joined, I built the first version myself, including the data cleaning, feature work and thresholds. The product question was how that signal could help the team decide its next intervention.
What I built
Spearheaded the 0→1 lifecycle for the career-support vertical, building the core platform, in-house ATS, alumni community, and placeability model from ground-up.
Before a data scientist joined, I built the first version of the placeability model myself: cleaning and tagging the data, engineering features and setting the operating thresholds. It achieved 74% recall.
Metrics
74% Recall
Delivered recall of 74% for candidate placeability prediction