Rajat User

Led product and stakeholder initiatives to fine-tune LLMs and built a web tool that automated dataset preparation and prompt workflows, improving departmental efficiency by 20% across 200+ employees.

Reframed physician-targeting into production-grade predictive models, reduced prediction and scoring time from 72 hours to 30 minutes, and delivered models adopted across therapeutic areas to improve targeting and sales effectiveness.

Built a real-time two-wheeler helmet violation detection system using YOLO, achieving class-wise mAP of 0.80 and 0.61 and reducing per-image processing time by 60%.

Developed credit risk and lead-scoring models that improved default prediction accuracy by 20% and increased sales conversion rates by 30%, enabling more precise credit decisions.

Developed object-detection and lead-tagging systems to automate hotel quality verification and lead allocation, reducing lead distribution turnaround from 7 days to 2 days and improving fraud detection.

Transitioned maintenance strategy to data-driven predictive maintenance and developed ML models to predict Remaining Useful Life and Mean Time Between Failures, improving operational reliability and uptime.

Activities and societies: Led mentorship program for PGDBA entrants; awarded perfect grade for PG dissertation.

Post Graduate Diploma in Business Analytics program focused on advanced analytics, ML, and data-driven decision making, completed with a dissertation jointly reviewed by IIM-C, ISI, and IIT-KGP.

Activities and societies: Conceptualised and executed a large student-led Science and Technology Festival; completed high-performing bachelor's dissertation (perfect score).

Bachelor of Technology in Mechanical Engineering with strong academic performance and a dissertation on catalytic combustion research published at the Asia Pacific Conference.