HEMANK JAIN

Classified one-year radon time series data into seismic and non-seismic zones using stacked machine learning models (GLM, LR, k-NN), significantly improving AUC from 0.83 to 0.93 across four rolling windows.

Identified and labeled full-day anomalies by integrating prediction confidence through an Automatic Anomaly Indication Function, triggering alerts at a confidence threshold of >= 0.55.

Enhanced classification accuracy from 2.9% to 23% by implementing hyperparameter tuning and class rebalancing techniques, effectively reducing seismic false negatives.

Developed a 5x5 meta-feature matrix from cross-validation predictions to efficiently aggregate signals and improve anomaly detection across dynamic time-series windows.

Curated 4 strategic workshops and 3 case sprints, achieving an average turnout of over 80 participants and increasing Product Management Club engagement by 30% within one semester.

Orchestrated speaker events featuring 5 industry leaders, including CPO of Airtel and Director of Domino's, to provide insights on Go-to-Market strategies, product roadmaps, and market fit.

Designed and executed the flagship APOGEE event focused on product strategy, attracting over 150 cross-campus attendees and doubling outreach compared to previous editions.

Directed 5 short films, including one pitched to Netflix, acted in 6 productions, and scripted 4 films and over 20 stage acts, performing before an audience of over 3,000 at APOGEE & OASIS.

Led a 50-member production team for 4 major acts and 5 flagship events, enhancing execution speed by 20% through structured delegation and weekly sprints.

Significantly scaled the Mime Club's digital reach by managing social media, achieving over 300,000 views on peak campaigns and increasing follower engagement by 2.4x.

Conducted over 60 internal workshops focused on scripting, performance, and team bonding, contributing to high team retention and consistent content output across semesters.

Enabled ₹5 Crore loan approvals by developing a comprehensive 10-year repayment projection model, incorporating DSCR 7.94, 4-year breakeven, and EMI ₹2.47L, utilizing advanced Excel-based modeling.

Reduced financial model error margin to less than 1% by rigorously validating over 15 key performance indicators (KPIs), including PAT, EBITDA, DSCR, and NWC, across three critical financial outcomes.

Designed and implemented a lender-facing dashboard that consolidated complex cash flows, repayment timelines, and credit risks, streamlining reporting and reducing loan approval time by 50%.