Looking for a job

Ayush Patnaik

@ayushpatnaik

Data Scientist at CRED who built lending optimisation systems that increased disbursals by 8% without reducing offer quality.

India

Message

At CRED, I built lending systems across underwriting, from risk-adjusted credit limits to offer allocation. My multi-objective allocation system increased disbursals by 8% without deterioration in offer quality, and my risk-adjusted limits model increased lending volume by 12% at comparable delinquency rates for sub-prime customers.

I also built production feature stores and monitoring pipelines, and developed models for fraud detection, reinforcement learning, and industry classification at Wells Fargo. Earlier, at Blackbuck, I built “Google Maps for trucks in India,” combining A* routing and machine learning in a web app for truck drivers.

Experience

Work history, roles, and key accomplishments

Current

Data Scientist L5

Current

CRED

Jan 2024 - Present (2 years 9 months)

Built risk-adjusted limits using double-machine-learning, optimized offer allocation, and developed portfolio optimization systems. Mentored data scientists and engineered AI agents for production support.

Machine Learning Double Machine Learning Optimization PySpark AI Agent Engineering Mentorship

Data Scientist L4

CRED

Jan 2022 - Jan 2024 (2 years)

Built multi-stage income and occupation models, expanded credit eligibility, and developed orthogonal risk discovery models. Led ML infrastructure and platform migrations.

Clustering NLP Tabular ML Feature Stores Databricks MLFlow AWS

Quantitative Analytics Specialist

Wells Fargo

Jan 2020 - Jan 2022 (2 years)

Built fraud detection models, developed workplace optimization models, and implemented reinforcement learning for personalized outreach. Created NLP classifiers for industry classification.

Fraud Detection Optimization Reinforcement Learning Q Learning NLP Neural Networks

Intern

Blackbuck

Jan 2019 - Dec 2019 (11 months)

Analyzed freight lanes and driver delays, built routing models combining A* with ML, and delivered a JavaScript web app for truck drivers.

A* Algorithm Machine Learning JavaScript Web Development

Intern

[24]7.ai

Jan 2018 - Dec 2018 (11 months)

Forecasted hourly chat demand using ARIMA and neural networks, and built linear programming models for agent staffing and service-level improvement.

ARIMA Neural Networks Linear programming Forecasting

Education

Degrees, certifications, and relevant coursework

Indian Institute of Technology, Kharagpur

B.Tech., Industrial & Systems Engineering

2015 - 2020

Activities and societies: Micro-specialisation: Entrepreneurship & Innovation

Pursued a dual degree program, earning a B.Tech. in Industrial & Systems Engineering and an M.Tech. in Industrial Engineering & Management, with a micro-specialisation in Entrepreneurship & Innovation.