Job Description:

This role is for someone who can take ideas from research and experimentation through to engineered, production-ready implementations. Youll combine data science/ML, full-stack development, and hands-on prototyping to deliver measurable outcomes for real-time, high-impact problems.

You’ll own delivery end-to-end: problem framing, data exploration, prototyping, experimentation, building robust engineering services and (where needed) user-facing components, deploying into pilot and production environments, and monitoring impact.

In this role you’ll:

1) Prototype, experiment, and validate

- Turn problem statements into testable hypotheses and rapid prototypes (models, features, rules + ML hybrids, decisioning logic).
- Design and run controlled experiments (A/B, champion/challenger, shadow mode, back-testing) to validate changes safely.
- Define success measures using business-relevant metrics such as detection rate, false positives, customer friction, operational impact, and latency, and translate results into clear recommendations.

2) Build production-grade ML and services

- Design, develop, and optimize ML models for fraud/scam detection across outbound payment journeys.
- Engineer and productionise features/labels from high-volume event streams and historical datasets, prioritizing stability and reproducibility.
- Implement ML solutions into POC, pilot and then production environments, including integration patterns, performance tuning, and operational readiness.
- Build supporting engineering components required to operate ML in production: APIs/services, batch/stream jobs, evaluation harnesses, model/version management, monitoring dashboards, and alerting.

3) Full-stack development to enable end-to-end delivery

- Build and maintain backend services and APIs (Python) that expose model decisions, features, and experimentation controls.
- Develop lightweight front-end/UI components where needed (e.g.,



internal tools for analysts/ops, experiment dashboards, model performance views), focusing on usability and clear decision explainability.
- Integrate solutions with enterprise systems (authN/authZ, logging, telemetry, secrets management), ensuring secure-by-design implementation.
- Apply strong engineering practices: clean architecture, code reviews, automated testing, CI/CD, and performance optimization.

4) Own the delivery lifecycle end-to-end

- Support all steps in the ML lifecycle: development, integration, testing, release, deployment, monitoring, incident support, and governance.
- Produce and maintain code repositories, artefacts, documentation, runbooks, and model documentation aligned to governance requirements.
- Partner closely with fraud strategy, operations, product, and technology colleagues to ensure solutions are practical, explainable, measurable, and deployable at pace.

To be successful in this role you should meet the following requirements:

Core experience (5 years)

- 5 years’ experience delivering end-to-end solutions spanning Data Science/ML and full-stack engineering, including hands-on production deployment and operational ownership.
- Demonstrated ability to move from R&D;/prototype engineered build production outcomes (not research-only work).

Data science / ML

- Strong background in applied data science (applied delivery is essential).
- Hands-on experience building and evaluating models.
- Strong Python skills with NumPy, Pandas, scikit-learn, TensorFlow,



and experience working with large-scale data processing frameworks such as Spark, Dask, and Hadoop.
- MLOps and production ML experience, including building and operating ML pipelines using tools such as MLflow (experiment tracking/model registry), Kubeflow and/or Vertex AI Pipelines (workflow orchestration), Airflow/Composer (scheduling), Docker and Kubernetes (containerization/orchestration), CI/CD (e.g., Jenkins/GitHub Actions/GitLab CI), feature stores (e.g., Feast/Vertex AI Feature Store), model monitoring/observability (e.g., Evidently, WhyLabs, Prometheus/Grafana, OpenTelemetry), data quality/testing (e.g., Great Expectations/Deequ), and secure deployment practices (secrets management, IAM, audit logging).
- Ability to explain model behavior, trade-offs, and results to technical and non-technical stakeholders.

Full-stack engineering

- Strong coding ability across backend and front-end (depth in one, working capability in the other).
- Backend/API development experience (REST), including integration with data stores and event/stream processing where relevant.
- Front-end experience building usable internal tools/dashboards (e.g., React/Angular/Vue or equivalent), with an understanding of secure UI patterns.
- Solid SQL and data access patterns; experience optimizing queries and designing data models is a plus.
- Engineering fundamentals: testing (unit/integration), CI/CD pipelines, observability (logs/metrics/traces), and performance tuning.

Cloud & platforms

- Cloud fundamentals; GCP preferred (or equivalent cloud experience with willingness to deepen GCP capability).
- Familiarity with containerization and runtime environments (e.g., Docker/Kubernetes) is beneficial.

Nice to have

- Exposure to quantum computing/quantum-inspired optimization and interest in applying advanced compute to cybersecurity and bank-wide use cases.

📌 Senior Consultant Specialist (Quantum,ML,Data,FSD) (Pune)
🏢 Purview Services
📍 Pune