Expert Product - Machine Learning

Expert: Expert Product - Machine Learning

Grade Level: L2

Location: Islamabad

Last date to apply: 21 September, 2026

What is Expert Product - Machine Learning?

You will be working with a team managing modelling and data-science layer behind JazzCash's lending and wealth products including credit risk, portfolio behavior, fraud, and customer analytics. This is a hands-on, build-heavy role: you will work directly on production data (200M+ loan records and counting), design the features and models that metricize the risk to check if credit is being assigned to the right customers at the right price. You will be accountable for how those models behave once they are live and will also be responsible for managing and retraining already deployed models. You will sit close to Product, Risk, and Data Engineering rather than behind a reporting queue.

What Expert Product - Machine Learning does?

1. Credit Risk & Lending Models

• Build, validate, and deploy application and behavioral scorecards for microlending - PD/LGD estimation, limit assignment, and repricing.
• Develop early-warning and collections-prioritization models using repayment behaviors, transaction, and telco signals.
• Engineer features from raw loan, wallet, and KYC data with explicit point-in-time correctness, so training data reflects what was actually knowable at decision time.

2. Portfolio & Risk Analytics

• Run survival analysis, vintage/cohort analysis, and band migration matrices across the full since-inception loan book.
• Support ECL provisioning models and stress scenarios in partnership with Risk and Finance.
• Investigate portfolio shifts - not just report them - and turn findings into changes to policy, cutoffs, or product design.

3. Production ML & Model Lifecycle

• Take models from notebook to production: containerized scoring, batch and near-real-time inference, orchestration, and rollback paths.
• Own monitoring for drift, stability (PSI/CSI), and performance decay, with alerting and a documented retraining trigger.
• Maintain model documentation, feature dictionaries, and explainability artefacts that satisfy internal audit and regulatory review.

4. Analytics Engineering & Data Pipelines

• Build and maintain analytical models in DBT on ClickHouse, and work with distributed PostgreSQL (Citus) and Teradata for source and serving layers.
• Write and tune SQL at scale - partitioning, join strategy, memory settings, and incremental/snapshot logic are part of the job, not someone else's problem.
• Orchestrate pipelines in Prefect (or equivalent) with proper failure handling, backfills, and data-quality checks.

5. Fraud, AML & Behavioral Analytics

• Develop anomaly-detection and network/graph models over transaction and identity data to surface fraud rings, mule accounts, and collusive behavior.
• Analyze customer transaction patterns to improve targeting, eligibility, and retention.

6. Experimentation & Measurement

• Design and read A/B tests, holdouts, and champion-challenger setups for scorecards, limits, pricing, and campaigns.
• Push back on underpowered or badly framed experiments, and quantify uplift honestly -including when the answer is "no effect".

7. Product Performance & Reporting

• Build KPI trackers and dashboards for lending, savings, and insurance verticals that people actually use to make decisions.
• Automate recurring reporting so analyst time goes to analysis rather than refreshes.

JazzCash is an equal opportunity employer. We celebrate, support, and thrive on diversity and are committed to creating an inclusive environment for all employees.

Requirements

What we are we looking for and what does it require to be an Expert Product - Machine Learning?

Required

• Bachelor's or master's in Data Science, Computer Science, Statistics, Applied Mathematics, Engineering, or a related quantitative field.
• 2-5 years in data science, machine learning, or advanced analytics - fintech, banking, telco, or digital financial services are strongly preferred.
• Strong Python (pandas, scikit-learn, and at least one gradient-boosting library) and genuinely strong SQL, including query optimization on large tables.
• Experience shipping models to production and keeping them alive afterwards. We don’t want candidates whose ML experience is restricted to Jupyter.
• Working knowledge of credit scoring and risk analytics: PD/LGD, vintage and roll-rate analysis, scorecard validation, and the standard model-performance metrics (like AUC, MAE).
• Comfort with version control, code review, and writing code other people have to maintain.

Strongly preferred

• Columnar/OLAP databases (ClickHouse or similar) and analytics engineering with DBT.
• PostgreSQL at scale, including distributed extensions such as Citus.
• Streaming and pipeline tooling: Kafka, Flink or Beam, Prefect/Airflow.
• Survival analysis, uplift modelling, or causal inference.
• Graph analytics for fraud/AML use cases.
• Model governance experience in a regulated environment including documentation, validation, and audit response.
• Dashboarding with Power BI, Tableau, or custom front-ends.

You will do well here if

• You are comfortable being wrong in front of people and correcting quickly.
• You would rather understand why a number moved than produce a prettier chart of it.
• You treat data quality and infrastructure as part of modelling, not a prerequisite someone else provides.

Benefits

Why Join JazzCash?

As one of the largest digital financial services providers in Pakistan, our objective is to continue to change the lives of our customers for the better.

Recognized as one of the leading employers in the country, JazzCash epitomizes the philosophy that each JazzCash employee is passionately living a better life every day, inspired and enabled by visionary leadership, a unique professional culture, a flourishing lifestyle, and continuous learning and development.

Our core values include qualities essential for a positive organizational culture - truthfully guiding entrepreneurial and innovative mindsets, harnessing professional and interpersonal collaboration, and fostering across-the-board customer obsession.

This is an opportunity for someone who wants to be part of something transformative, someone who can play a critical role in driving our success. Together, we can empower millions more with the tools necessary to progress in an increasingly digital economy.