ML Engineer

The role:

As a Senior ML Engineer, you'll play a central role in building the data science behind the products — from framing new fraud, risk, and intelligence problems, to designing and deploying ML models at scale, to helping our enterprise customers and go-to-market teams understand and act on the insights we generate.

What you’ll do:

• Design, build, and continuously improve the ML models that power our risk and intelligence products, and take ownership of new signals as they get scoped.
• Take a loosely defined business or customer problem and break it into a clear data problem, articulating value, impact, and complexity before proposing a solution.
• Build anomaly detection, fraud, and risk-modeling approaches — including network/graph-based methods — that keep our signals accurate and resistant to adversarial behaviour.
• Own model development, deployment, and monitoring end-to-end, partnering with ML/data engineers on scalability, reliability, cost, and dashboards/alerting.
• Design and run experiments (A/B tests, offline customer POCs) to validate new signals before they roll into production.
• Manage and analyse large, multi-country datasets, ensuring data integrity, consistency, and compliance throughout.
• Partner cross-functionally with Product, Engineering, Legal, and GTM/Sales to scope, prioritise, and ship data products on time, acting as a trusted advisor on what the data can and can't responsibly support.

What you bring in:

• 5+ years of experience designing, building, and deploying ML models at scale, ideally including risk/fraud, propensity, or behavioural/network scoring use cases.
• Strong grounding in applied machine learning: classification, anomaly detection, propensity/scoring models, clustering, and time-series/drift monitoring.
• Exposure to graph-based analysis or graph ML (network embeddings, community detection, link prediction) is a plus.
• Hands-on experience taking models from research/experimentation into production — comfortable owning scalability, reliability, and monitoring, not just model accuracy.
• Working knowledge of NLP and LLM-based techniques (prompting, summarisation, fine-tuning) — useful for customer-facing AI insights and on-device text/SMS signal extraction.
• Proficiency in Python and the ML/data stack: Pandas, NumPy, Scikit-learn, TensorFlow or PyTorch; comfortable with Hugging Face Transformers where relevant.
• Strong SQL skills and experience with large-scale data processing (BigQuery, Spark/PySpark, Hive/Kafka ecosystem).
• Familiarity with database modelling and data warehousing principles.
• Ability to design, run, and interpret experiments and statistical tests to validate model and business impact.
• Strong communication skills — able to explain model output and trade-offs to both engineering peers and non-technical enterprise stakeholders.
• Comfort operating with a strong privacy/compliance mindset — Truecaller's data products are built on abstracted, privacy-safe signals, and you'll need to reason carefully about what can and can't be derived or stored.

It would be great if you also have:

• Experience with graph databases (e.g. Neo4j) or large-scale graph processing frameworks.
• Experience with ML lifecycle tools (Kubeflow, MLflow) and on-device/edge ML deployment.
• Familiarity with Google Cloud Platform (GCP) and BigQuery.
• Prior experience in fraud/risk scoring, credit/alternate-data, or contact-centre/dialer analytics domains.
• Experience working with data resellers, credit bureaus, or enterprise data partners on integrating and explaining model output.