Senior Data Analyst — Risk & Business Analytics

Keka Inc.; Bengaluru; Hybrid; INR 1,200,000 - 1,800,000; Full time

Job summary

InPrime Finserv in Bengaluru invites a Senior Data Analyst — Risk & Business Analytics to own end-to-end analyses for credit, product and business teams. You will craft questions, data methods, conclusions and recommendations with clarity for senior leadership.

You will own dashboards in QuickSight/Superset, design fact/dimension models, and partner with data engineering to ensure trusted datasets. The role sits in a small team and reports to the AVP, Data Science & Analytics.

Qualifications

• 3+ years in analytics, preferably in lending/fintech.
• Advanced SQL with window functions, CTEs, cohorts.
• Dimensional modelling with grain and surrogate keys.
• Python for analysis using pandas and statsmodels/sklearn.
• Strong experience with BI tools (QuickSight/Superset and others).
• Statistical judgement and clarity in definitions.

Responsibilities

• Own portfolio risk analytics including curves and delinquency trends.
• Backtest credit policy changes and monitor impact.
• Lead business analytics: funnel, productivity, disbursement mix.
• Measure policy and product impact with cohorts and tests.
• Design analytical data models and partner with data engineering.
• Build self-serve dashboards backed by certified datasets and consistent metrics.
• Ensure data quality and fix root causes.

Skills

Advanced SQL

Dimensional modelling

Python for analysis

BI tools

Statistical judgement

Clear writing

Analytical leadership

Tools

QuickSight

Superset

Tableau

Power BI

Looker

Metabase

Job description

Senior Data Analyst — Risk & Business Analytics

3 + Years

HSR HO

Full-Time

Senior Data Analyst — Risk & Business Analytics

InPrime Finserv |Bangalore | Full-time| Hybrid — 3 days/week in office

3+ years experience |Reports to AVP, Data Science & Analytics

About the role

You will be the analyst the credit, product and business teams come to when they need a number they can act on.

Reporting is part of this job — you will own the dashboards the business runs on. But you are not a request queue. You will own analyses end to end — question, data, method, conclusion, recommendation — and be expected to explain and stand behind your conclusions with senior leadership and the founders.

You will report to the AVP, Data Science & Analytics and sit on a team of four data scientists and analysts. No layers in between: the person setting the analytics agenda is the person you work with directly.

What you'll own

• Portfolio risk analytics: vintage/cohort curves, roll rates, DPD migration, bounce and collection efficiency, delinquency and loss trends by segment.
• Credit policy analytics: backtest and size BRE changes before launch; measure approval-rate, bad-rate and segment impact; monitor cutoffs, overrides and swap sets after launch.
• Business analytics: funnel conversion, branch/RM productivity, disbursement mix, yield, portfolio composition and operational bottlenecks.
• Experimentation & impact measurement: determine whether policy, product and process changes actually worked, using appropriate cohorts, holdouts, matched comparisons or statistical tests.
• Analytical data models: design fact/dimension tables, loan-level daily panels and analytical marts. You own the analytical design — grain, keys, dimensions and metric logic — and partner with data engineering on upstream data and productionisation.
• Self-serve reporting: build durable, documented dashboards in QuickSight and Superset backed by certified datasets and consistent metric definitions.
• Data quality: notice when a number looks wrong, trace it to source and get the underlying issue fixed rather than patching individual queries.

Data agent semantic layer

We have built an internal LLM-based data agent so the business can self-serve ad-hoc questions. You will own the analytical layer it reasons over — certified datasets, valid join paths, metric definitions and business terminology — test it against real questions, investigate wrong answers and improve the underlying semantics.

You do not need prior LLM experience. We would rather hire strong analytical judgement and teach the LLM side.

The scope above is broad by design, and we have a view on sequence:

• Own the risk and portfolio reporting layer end to end — certified datasets, consistent definitions, dashboards credit and collections rely on daily.
• Build out the business and funnel reporting layer — lead-to-disbursement funnel with stage-wise drop-offs and turnaround times, branch and RM productivity, disbursement mix and portfolio composition — so product, business and the founders stop asking us for these numbers.
• Become the analytical partner to credit on policy changes, from pre-launch sizing through post-launch monitoring.
• Take ownership of the data agent's semantic layer and start cutting the ad-hoc queue.

Deeper experimentation and impact measurement build from there. We will revisit this with you in the first month rather than hand it down.

What you need

• 3+ years in analytics, ideally at a lender, fintech, NBFC, bank or consumer-internet company. Lending experience is a strong plus.
• Advanced SQL: window functions, complex CTEs, temporal logic, cohorts and longitudinal/panel data.
• Dimensional modelling: grain, fact/dimension tables, surrogate keys, slowly changing dimensions and additivity.
• Python for analysis: pandas, numpy and basic statsmodels/sklearn.
• Strong experience with a BI tool. We use QuickSight and Superset; Tableau, Power BI, Looker, Metabase etc. transfer fine.
• Statistical judgement: cohort maturity, confidence intervals, mix effects, and why before/after movement alone does not establish impact.
• Rigour about definitions and clarity in writing. You should be irritated by two dashboards showing different numbers for the same metric and fix it at the definition, not in the query. And you should be able to take a loosely defined business question, turn it into a testable one, and write up the answer in a page.

Nice to have

• Bureau, bank statement, collections or alternate-data experience
• AWS analytics stack — Athena, Glue, S3, Bedrock
• dbt or similar transformation frameworks
• Scorecards, WoE/IV, logistic regression, ECL / Ind AS 109

How we work

Small team, high autonomy, low process overhead. Based in Bangalore, 3 days per week from office.

You will work directly with credit, product, engineering, compliance and the founders, and day to day with the data scientists on the team.

Your analysis will change what gets shipped — which means we would rather you take an extra day and be right than ship fast and be wrong.