Role : Data Scientist - Core ML & Deep Learning

Build AI that moves from notebook to production - and transforms insurance outcomes for better customer experience.

What you'll own :

- Shape the problem : Work with business and product teams to define the decision, target, prediction grain, outcome window, leakage controls and cost of errors.

- Create signal from data : Use Python, SQL/PySpark to explore large datasets, engineer useful features and identify missingness, imbalance, drift and segment-level behaviour.

- Build models that earn trust : Develop strong baselines and fit-for-purpose ML/DL solutions across tabular data, NLP, documents or images when they improve the outcome.

- Prove what works : Use time/cohort-aware validation, relevant metrics, threshold optimisation, error slicing, robustness checks and explainability such as SHAP.

- Ship, don't shelf : Package models for batch or API inference; integrate testing, Git, Docker, model registries and CI/CD with engineering partners.

- Keep improving : Monitor model and business performance, drift, latency and failures; define retraining, rollback and human-review paths.

What you'll bring :

- Hands-on experience : 1 - 5 years in data science, ML or applied AI, including at least one model moved beyond a proof of concept.

- Strong foundations : Python, SQL, pandas, NumPy, scikit-learn, statistics, feature engineering, imbalanced learning and model evaluation.

- Modern model toolkit : XGBoost, LightGBM or CatBoost, plus PyTorch or TensorFlow for relevant deep-learning, NLP or computer-vision use cases.

- Production mindset : Comfort with APIs or batch services, testing, Git, Docker and cloud/MLOps fundamentals; AWS/SageMaker, MLflow or Spark is a plus.

- Clear thinking : You can explain assumptions, trade-offs and limitations simply, and collaborate well with domain experts, product managers and engineers.

What will make you stand out :

- End-to-end proof : A model you personally moved from problem framing and validation into a real pilot or production workflow.

- Experimental rigour : A clear explanation of baselines, leakage prevention, metric choice, threshold trade-offs and segment-level error analysis.

- Ownership under ambiguity : Examples of turning an unclear business ask into a measurable ML problem and influencing stakeholders with evidence.

Why this opportunity stands out :

- Real-world scale : Work on decisions that can improve the experience of millions of policyholders - not toy datasets or isolated experiments.

- End-to-end ownership : Stay close to the journey from discovery and modelling through deployment, adoption and measurable business impact.

- A builder's environment : Experiment fast, learn from strong peers and help shape the engineering standards of Star Health's growing Enterprise AI team.

- Career acceleration : Build domain depth, production judgement and stakeholder influence while creating a portfolio of work with visible enterprise impact.

What your first six months can look like :

- Learn the landscape : Understand the decision journey, data, users and success metrics; reproduce the baseline and agree the production quality gates.

- Ship a meaningful release : Move a priority model into a controlled pilot or production path with explainability, monitoring and documented operating controls.

- Show measurable value : Improve a business or customer outcome and turn what worked into reusable features, pipelines or modelling standards for the next use case.

Your impact at Star Health :

Build for differentiation : underwriting risk and straight-through processing; Fraud Wastage & Abuse models for claims triage, severity and fraud; renewal/lapse propensity; document intelligence; service and demand forecasting.

Success looks like : a model is accurate, explainable, production-ready, monitored and actively used to make a better business or customer decision.

READY TO BUILD?

Bring your curiosity, your GitHub or portfolio if you have one, and a story about something you shipped. You do not need to tick every box - show us how you learn and solve.