Open to opportunities

Sheetal Kumar Jain

@sheetalkumarjain

Message

I lead data science teams delivering measurable gains across risk and operations.

India

Message

At Navi Limited, I lead a 10+ member Data Science team across debt management, ads, credit risk, and investments. I redesigned allocation prioritization to improve D30 by 26 bps and scaled ads data science coverage from 20% to 50% with an 18% CPM uplift.

Previously, I built ML infrastructure and real-time intent systems at GoKwik, fraud and identity models at Gojek, and credit risk models at American Express. My work spans machine learning, deep learning, NLP, GenAI, risk scoring, and production systems that improve precision, reduce costs, and drive business outcomes.

Experience

Work history, roles, and key accomplishments

Current

Data Science Manager

Current

Navi Limited

May 2025 - Present (1 year 4 months)

Leading a 10+ member Data Science team across Debt Management, Ads, Credit Risk, and Investments, driving cross-functional strategy execution and measurable business impact. Improved D30, Cycle 1–4 & 30+ EMI CE by 26 bps, 16 bps and 1.5pp via allocation-prioritization redesign.

Engineering Manager, Data & Applied Science

GoKwik

Jul 2022 - May 2025 (2 years 10 months)

Led development of a Real-time Pre-Order Intent Model, increasing precision by 50% and recall by 100%, unifying 20+ models. Built and scaled reliable end-to-end ML infrastructure including monitoring, feature store, alerts, dashboards, and auto-refresh pipelines.

Data Scientist 2, Fraud and Identity

Gojek

May 2021 - Jul 2022 (1 year 2 months)

Built the first Customer Risk Scoring (CRS), reducing promo abuse and delivering ~$3.8M monthly savings. Engineered a LGBM based Pre-Login Risk Score (PLS) for ATO prevention, improving detection accuracy by 70%.

Machine Learning Deep Learning NLP Python LGTM TF IDF CNN EfficientNet Fraud Detection Risk Scoring Account Takeover (ATO) Prevention

Assistant Manager, MLDS Research

American Express

Jul 2018 - May 2021 (2 years 10 months)

Built CGAN-based data augmentation framework, driving 3.64% model lift; presented at ODSC 2019 and scaled across portfolios. Architected multiple DL credit risk models (Spend-Payment RNN, RNN LOTM, Multi-head ANN) for default prediction.