MobiKwik
Website:
mobikwik.com
Job details:
MobiKwik is one of India's leading fintech platforms — operating at the intersection of payments, lending, and wealth management for millions of consumers and merchants. Data science sits at the core of how we underwrite credit, detect fraud, personalise products, and grow responsibly. We are looking for a Head of Data Science who will own the function — from credit risk and fraud ML to experimentation, personalisation, and the long-term AI roadmap. This is a senior leadership role with direct P&L impact. You will work closely with the Head of Data Analytics & Products, business heads, and the executive team to make data science a genuine competitive advantage for MobiKwik.
What You will Do
Credit Risk & Lending Science
• Own the end-to-end credit underwriting pipeline — from bureau-based scorecards to
behavioural and alternate data models for thin-file and NTC customers.
• Build and monitor models across the loan lifecycle: origination scoring, limit management,
early warning systems, collections propensity, and recovery prioritisation.
• Maintain model health through continuous monitoring of score distributions, Gini coefficients,
vintage curves, and delinquency trends across DPD buckets.
• Collaborate with risk, credit policy, and collections teams to close the loop between model
output and business outcomes.
Fraud Detection & Payments Intelligence
• Lead the development of real-time fraud detection models across payment flows — account
takeover, transaction fraud, merchant fraud, and synthetic identity.
• Work with product and engineering on model serving infrastructure: latency constraints,
feature pipelines, and model-in-the-loop decisioning.
Personalisation, Recommendations
• Build propensity and recommendation models for cross-sell and upsell across payments
AI Strategy & Platform
• Define and own MobiKwiks multi-year AI roadmap — from near-term model improvements to strategic bets on GenAI, foundation models, and agentic workflows.
• Evaluate and drive adoption of emerging AI capabilities in areas like collections automation,
customer service, document processing, and code productivity for the data team.
Team Building & Stakeholder Influence
• Hire, develop, and retain a high-performing team of data scientists and ML engineers — build
a culture of ownership, intellectual rigour, and business proximity.
• Structure the team to balance depth (specialists in credit, fraud, NLP) with breadth (full-stack
scientists who can own problems end-to-end).
Non-Negotiables
Hands-On Execution
You are not a manager of managers who hasn't touched code in five years. You can review model code, debug a feature pipeline, question a loss function, and get into a Jupyter notebook when it matters. You hold the team to high technical standards because you've done the work yourself.
Long-Term AI Vision
You think beyond the next model release. You have a point of view on where AI in fintech is heading — LLMs in credit, real-time personalisation, agentic workflows — and can translate that into a sequenced roadmap that makes sense for a company at MobiKwik's scale and stage.
Strategic & First-Principles Thinking
You don't default to what the last company did. You ask why before how. You can structure ambiguous problems from scratch, identify the key lever in a complex system, and prioritise ruthlessly when resources are constrained.
Stakeholder Influence & Team Building
You can walk into a credit committee with a risk head who is sceptical of black-box models and leave with alignment. You can write a one-pager for the CEO that makes a technical case for investment. And you build teams where people grow, stay, and do the best work of their careers.
What You Bring
Experience
• 8–14 years in data science or ML, with at least 3 years in a leadership role owning a science
team end-to-end.
• Proven track record in a fintech, NBFC, or bank environment — credit risk, fraud, or payments
ML is a strong signal.
• Experience building and deploying models at scale in production — not just prototypes.
• Track record of influencing business outcomes with data science, not just delivering models.
Technical Depth
• Strong foundations in statistical modelling and ML: gradient boosting, survival models, neural
networks, NLP.
• Hands-on proficiency in Python (scikit-learn, XGBoost, AutoML, TensorFlow) and SQL.
• Experience with MLOps: model serving, feature stores, drift monitoring, retraining pipelines.
• Comfortable with large-scale data processing (Spark, BigQuery, or equivalent).
• Familiarity with bureau data ecosystems (CIBIL, Experian, CRIF) and alternate data sources
is a plus.
• Understanding of LLMs, embedding models, and where GenAI is genuinely useful vs.
overhyped.
Domain Knowledge
• Deep understanding of credit underwriting mechanics: scorecards, cutoff strategy, reject
inference, vintage analysis.
• Familiarity with the RBI's digital lending framework, FLDG regulations, and fair lending
expectations for model-based decisions.
• Awareness of India's data and credit ecosystem — UPI transaction data, AA framework, GST data, NACH mandates as ML signals.
What Great Looks Like in This Role
In the first 6 months, a great hire will:
• Have a complete map of every model in production — what it does, how it performs, and
what's at risk.
• Have shipped at least one meaningful improvement: a model re-trained with new data
sources, a feature that moved a key metric, or a new monitoring system that caught
something before it became a problem.
• Have earned the trust of the risk, product, and business teams — they come to you with
problems, not just ask for deliverables.
• Have a 12-month science roadmap that's been reviewed by leadership and is being actively
resourced.
In the first 12–18 months:
• The data science team has clear career ladders, measurable OKRs, and a culture of rigour
and ownership.
• MobiKwiks credit models are measurably better: approval rates up, NPA stable or down,
model explainability improved for audits.
• At least one AI initiative beyond classical ML is live or in advanced pilots — be it an LLM-
based tool, an agentic workflow, or a real-time personalisation layer.
• You have a strong hiring pipeline and have upgraded team quality at least one tier.
Click on Apply to know more.