Data Scientist

Snapmint; Bengaluru; On-site; INR 1,200,000 - 2,400,000; Full time

Job summary

Snapmint is hiring a Data Scientist to own end-to-end credit underwriting models for a high-scale fintech product. The role covers model development, real-time deployment, monitoring, and continuous improvement, directly influencing approval rates, fraud, credit losses, and GMV.

Responsibilities include building credit risk models, designing underwriting strategies for New-to-Credit users, and balancing risk with conversion in checkout flows.

Qualifications

• 1-3 years of experience in Credit Risk / Lending / Cards.
• Hands-on experience with Python & SQL.
• Experience with credit bureau data (CIBIL / Experian / Equifax).

Responsibilities

• Build and maintain credit risk models using Logistic Regression, XGBoost/LightGBM and deep learning.
• Design underwriting strategies for New-to-Credit users using alternative & behavioural data.
• Balance credit risk, fraud risk, and conversion in checkout flows.
• Design real-time features with latency SLAs under 100 ms.
• Ensure training–serving feature parity using feature stores or custom pipelines.
• Partner with Risk teams to define score cut-offs, bands and guardrails.

Skills

Python

SQL

Credit Risk

XGBoost

LightGBM

Deep Learning

Job description

We are hiring a Data Scientist to own end-to-end credit underwriting models for a high-scale fintech product. This role spans model development, real-time deployment, monitoring, and continuous improvement, directly impacting approval rates, fraud, credit losses, and GMV.

What You’ll Do

• Build and maintain credit risk models using Logistic Regression, XGBoost / LightGBM and deep learning
• Design underwriting strategies for NTC (New-to-Credit) users using alternative & behavioural data
• Balance credit risk, fraud risk, and conversion in checkout flows
• Design real-time features with strict latency SLAs (<100 ms):
• Velocity & frequency features
• Rolling window aggregates
• Behavioural & transactional signals
• Ensure training–serving feature parity using feature stores or custom pipelines
• Partner with Risk teams to define:
• Score cut-offs, bands, and guardrails
• Policy rules layered over ML models

Experimentation & Business Impact:

• Measure impact on:
• Approval rate

What We’re Looking For:

• 1-3 years experience in Credit Risk / Lending / Cards
• Strong hands-on experience with Python & SQL
• Experience with credit bureau data (CIBIL / Experian / Equifax)