×
0 followers0 following
Jay Vardhan Sharma
Rajasthan, India
I'm currently a fourth year student at BIT, Mesra. I'm pursuing BTech in CSE.0 followers0 following
- About
- High School
- Bachelor's
- Master's
- Target
- TOEFL
- IELTS
- Duolingo
- LSAT
- GRE
- MCAT
- GMAT
- SAT
- ACT
Male
Rajasthan, India
Work ExperienceAI/ML & Data Analyst Intern
Office of the Principal Accountant General (A&E) - Internship
May 28, 2025 - Jul 28, 2025 · 2 mos
Rajasthan, India
◦ Developed an ML-based fraud detection system for government financial audit data using Isolation
Forest and statistical analysis.
◦ Engineered 40+ domain-specific features, including election proximity flags, GST/IT deduction flags,
rolling transaction windows, and other temporal, statistical, interaction, and cyclic features.
◦ Integrated Z-score logic, duplication checks, and rolling window mechanisms to improve fraud
detection accuracy.
◦ Visualized anomalous spending behavior using Seaborn and Matplotlib to support audit decisions.
◦ Optimized fraud detection by replacing manual full-dataset scanning with an ML model that targets
a 4–5% high-risk subset, boosting audit efficiency.
◦ Mentored by Shri Pravindra Yadav, Principal Accountant General (A&E), Jaipur
ProjectsOnline Educational Platform
Role - Full Stack Developer
for Personal
Sep 2, 2024 - Sep 12, 2024
Built a scalable educational platform offering free learning resources, mentorship, and structured career guidance.
Developed an AI-powered adaptive quiz engine to assess performance and personalize content delivery.
Implemented modular backend logic with interactive UI to enhance usability and system maintainability.
Research PapersMoving Beyond Black Boxes with Interpretable Anomaly Detection in Government Finances
Dec 12, 2025 - Springer
VolunteeringData Analytics & Financial Transparency Associate
at Institute of Community Development (ICD), Jaipur
Oct 6, 2025 - Nov 26, 2025
Children
Designed a lightweight, explainable financial analysis framework using rolling statistics.
Implemented duplicate and redundancy detection via fuzzy matching, uncovering reporting inconsistencies.
Translated analytical findings into clear, non-technical explanations to support ethical financial decisions.