Ajay Tiwari

I’m a B.Tech CSE (AI) student graduate June 2026, and I focus on turning messy data into risk-ready insights. I’ve built projects around credit and fraud risk analytics, customer risk segmentation, and data governance—always aiming for models and dashboards that stakeholders can trust.

In my payment fraud project, I sampled 55,554 transactions from the PaySim dataset and validated data governance with 0 duplicates and 0 nulls in SQL and Python. I engineered a value_category risk feature to reveal higher fraud probability in high-value transactions, and I identified fraud concentration patterns across TRANSFER and CASH_OUT.

For customer outcomes, I built a churn and cross-sell segmentation approach on 8,000 users where single-metric models failed to separate groups. I delivered a 3-tier segmentation with mapped retention actions, and I used subscription × device analysis to pinpoint churn drivers like Student + Mobile.

Recently, I completed a two-month LLM post-training internship at Ethara AI, assessing 300+ model outputs for instruction following, truthfulness, and verbosity across cross-functional teams. I’ve also strengthened my workflow with Power BI/Tableau dashboards, KPI reporting, trend analysis, and data quality checks—supported by an INSPIRE scholarship and a Start Up Arena finalist recognition.