Looking for a job

Kunal Maurya

@kunalmaurya

Senior Data Scientist at FairMoney who built a multi-market PD calibration layer for credit decisions.

At FairMoney, I built its first reusable multi-market calibration layer, aligning expected PD with actual defaults across Nigeria, Zambia, and Uganda. It enabled model refreshes without policy-cutoff changes and is expected to reduce rollout time.

I developed acquisition and behavioural risk models using bureau, application, third-party, and repayment signals, reducing the FPD30 amount rate by 5 percentage points and saving approximately EUR 25K. I also designed PD-based pricing tests that improved uptake from 70% to 85% while keeping risk stable and gross profit within target.

Earlier at Paytm, I built merchant credit-risk and PD models and bureau feature-store pipelines for underwriting and exposure eligibility. At Capital Float, I developed risk and collections scores, including a customer management score whose KPI dashboards contributed to a 30% reduction in 75th-percentile TAT.

Experience

Work history, roles, and key accomplishments

Current

Data Scientist - Modelling, Calibration & Experimentation

Current

FairMoney

May 2022 - Present (4 years 5 months)

Built FairMoney's first reusable multi-market calibration layer across Nigeria, Zambia and Uganda, aligning expected PD with actual default rates. Developed acquisition and behavioural risk models, price-sensitive credit risk models, and PD-based pricing A/B tests, improving uptake from 70% to 85% while keeping risk stable.

Python PySpark SQL BigQuery Logistic Regression LightGBM PD Modeling Calibration A B Testing Pricing Strategy Risk Monitoring Tableau Early Warning System (EWS)Feature Engineering MLOps Git Docker Airflow

Data Scientist - Merchant Lending Risk

Paytm

Apr 2019 - May 2022 (3 years 1 month)

Built and scaled Paytm's merchant credit-risk and PD models using transaction behaviour, business activity, repayment history, merchant profile and bureau variables. Developed merchant ability-to-pay simulation algorithms and bureau feature-store pipelines for exposure eligibility and risk analytics.

Python SQL PySpark Feature Engineering Bureau Data PD Modeling Risk Analytics Simulation BigQuery

Data Scientist

Indifi Technologies

Oct 2018 - Mar 2019 (5 months)

Created bank-statement transaction-description clustering and n-gram features to improve financial-behaviour signals for risk modelling.

Python NLP Feature Engineering Risk Modelling

Decision Scientist

Capital Float

Sep 2016 - Aug 2018 (1 year 11 months)

Built propensity-to-bounce collections model achieving AUC 77% and KS 38%. Developed bureau risk score for business-loan underwriting and customer management score for renewals, enhancements and cross-sell, achieving AUC 82% and KS 40%.