Head of Data Science, Transaction Analytics (Client Role)

Head of Data Science, Transaction Analytics (Client Role)

Experience: 8 to 10 years | Employment type: Full-time | Location: India, remote

About us

Jash Data Sciences (JashDS) is a Data + AI company based in Pune. We build data platforms, AI applications and agents for clients ranging from startups to enterprises, from the data foundation through to production.

Our culture: we value intellectual curiosity, continuous learning and collaborative problem-solving.

About the role

You will lead data science for one of our clients, a transaction intelligence fintech serving banks and commerce platforms across Asia-Pacific. Their platform already cleans, categorises and resolves bank, card, payment and purchase transactions to real merchants, in production across multiple markets.

The next layer is yours: models that anticipate customer behaviour rather than describe it, and the representation of a customer that every downstream model starts from, built for banks first and for commerce platforms alongside.

This is a senior, hands-on role. You frame the problem with the client's customers, design the approach, build the first version yourself and defend it in front of their data science and risk teams. You will work directly with the client's co-founders and engineering lead, and lead a small data science team that you will grow.

What will you be doing?

• Build propensity, next best action, early-warning and behavioural risk models on transaction histories, validated to a standard that banks' personalisation and risk functions accept
• Build the customer representation that downstream models start from
• Take the current transaction intelligence product to a standard where its output is trusted without anyone re-checking it
• Act as the senior counterpart to banks' and platforms' data science and risk teams: scope pilots, present results and benchmark our models against theirs
• Hire and coach the data science team, and decide with the founders which client questions become products

What do we need from you?

• 8 to 10 years in data science (not analytics, BI or risk policy)
• Predictive models on behavioural data (gradient-boosted, propensity or uplift models) that you built yourself and took to production, and can describe in detail
• Transaction data (account, card, payment or purchase) as the primary input to those models, gained at a bank or a payments fintech
• Credit risk models built and validated for production. Policy or rule-writing alone is not enough
• Still hands-on in Python and SQL on large datasets: feature pipelines on hundreds of millions of rows, model code you would put in front of a reviewer, and the means to reproduce it
• At least one pipeline or model that other people continued to run after you moved on
• Experience presenting to a bank's data science or risk leadership, and comfortable framing an open problem before solving it

Good to have

• Sequence, embedding or representation models on transaction or event data
• Deep learning applied where it has earned its place over simpler models
• A model taken through a bank's model-risk or validation process
• Using LLMs for data labelling, enrichment or extraction at scale

Qualification

• Bachelor's or Master's degree in Computer Science, Statistics, Mathematics, Data Science or a related quantitative field

What We Offer

• Model clean, enriched, transaction-level data across multiple Asia-Pacific markets
• Work directly with the client's founders, with your models in front of banks and commerce platforms across the region
• Shape what the product offers next
• Competitive compensation