About the role
Job Description
AI/ML Engineer / Data Scientist AI Product Enablement
Grade: Senior Director
Experience: 12–18 Years
Location: Bengaluru
Role Overview
We are looking for a highly experienced AI/ML Engineer / Data Scientist to drive AI product enablement across Comviva's product ecosystem, including mobile money, digital wallet, digital banking and financial services platforms.
The role will involve working closely with Product, Engineering, Data, and Business teams to identify AI/ML opportunities, design and develop production-grade AI solutions, and embed intelligence into products and platforms. A core part of the role is the hands-on creation of reusable machine learning models for high-value financial services use cases, such as fraud and risk detection, cash and liquidity projection, and credit scoring/rating, delivered as configurable product capabilities that can be adapted across customer deployments.
The ideal candidate should have strong hands-on experience in AI/ML, Generative AI, LLMs, machine learning systems, data science, and product engineering, along with the ability to translate business/product requirements into scalable AI solutions and to take models from data exploration all the way to monitored, governed production systems.
Key Responsibilities
Strategy & Opportunity Identification
- Define and drive the AI/ML technical roadmap for product and platform capabilities.
- Identify and prioritise opportunities to embed AI/ML and Generative AI into existing and new products, based on business value, data readiness, feasibility and regulatory considerations.
- Evaluate emerging AI technologies, frameworks, and models and assess their applicability to Comviva products.
Model Creation & Delivery
- Own the end-to-end lifecycle of ML models: problem framing, data exploration, feature engineering, training, tuning, validation, deployment, and continuous improvement.
- Design, build and deploy production-grade models for key use cases including fraud detection and anomaly detection, cash and liquidity forecasting, credit scoring and risk rating, customer analytics and recommendations.
- Build pre-trained, configurable model templates and feature libraries that can be tuned to each customer deployment (operators, banks, MFIs, fintechs), including approaches for limited-data and cold-start situations.
- Design real-time, low-latency inference for scoring in the transaction path, as well as batch and near-real-time scoring for forecasting, risk and analytics workloads.
- Work on Generative AI, LLMs, RAG, agentic workflows, NLP, predictive analytics, recommendation systems, anomaly detection, and other relevant AI use cases.
Architecture, MLOps & Governance
- Design AI/ML architectures covering data pipelines, feature engineering, model development, deployment, monitoring, and continuous improvement.
- Develop and implement MLOps practices for model deployment, monitoring, versioning, retraining, and governance, including drift and performance monitoring in production.
- Establish model governance practices: explainability, bias and fairness testing, validation, documentation and audit trails to meet regulator, auditor and customer risk-management expectations.
- Ensure privacy and security of data and models, including PII protection, data minimisation and compliance with applicable data protection and financial regulations.
Collaboration & Leadership
- Collaborate with Product Managers, Architects, Engineering teams, and business stakeholders to convert product requirements into AI-driven solutions.
- Provide technical leadership: set standards, review designs and code, and mentor data scientists and ML engineers.
Required Technical Skills
AI / Machine Learning
- Strong expertise in Machine Learning and Deep Learning.
- Experience with supervised and unsupervised learning; NLP, recommendation systems, predictive analytics, anomaly detection, and classification.
- Proven experience building and deploying production-grade ML models at scale for fraud/anomaly detection, time-series forecasting and credit risk scoring or comparable use cases.
- Strong command of tabular modelling (gradient boosting, ensembles) and time-series forecasting (statistical, ML and deep learning approaches).
- Hands-on handling of real-world data challenges: highly imbalanced classes, cost-sensitive learning, delayed or noisy labels, selection bias, concept drift, and feature engineering from transaction and behavioural data.
- Rigorous model validation along with Explainable and responsible AI
- Graph analytics / graph ML for network-based fraud detection is desirable.
Generative AI / LLM
- Strong understanding of Generative AI and Large Language Models (LLMs).
- Experience with RAG, embeddings, vector databases, prompt engineering, fine-tuning, and model evaluation.
- Exposure to agentic architectures, tool use, guardrails, hallucination control, and LLM security (such as prompt injection and data leakage).
Programming
- Strong proficiency in Python and the ML ecosystem (scikit-learn, XGBoost/LightGBM, PyTorch or TensorFlow, statsmodels/Prophet or similar).
- Good understanding of Java/Scala or another programming language is desirable, for integrating models with high-throughput, JVM-based transaction platforms.
- Strong knowledge of data structures, algorithms, and software engineering practices.
Data & Big Data
- Strong SQL skills; experience with relational and NoSQL databases.
- Exposure to Spark/PySpark, Kafka, Hadoop or other distributed data-processing technologies, including streaming platforms (Kafka Streams, Flink or Spark Structured Streaming) for real-time features.
- Experience designing data pipelines and feature engineering at scale for AI/ML workloads, with attention to data quality and lineage.
Cloud & MLOps
- Experience with one or more cloud platforms: AWS, Microsoft Azure, Google Cloud Platform; familiarity with private-cloud and on-premise deployments commonly required by telecom and banking customers.
- Exposure to Docker / Kubernetes, MLflow, CI/CD, model deployment and monitoring, feature stores, and MLOps frameworks and practices.
- Experience with model serving and optimisation for low-latency inference (such as ONNX, KServe, Triton or similar), and with drift detection and automated retraining.
Domain Knowledge (Preferred)
- Understanding of payments, mobile money, digital wallets and digital banking, including fraud typologies and AML/CFT concepts.
- Familiarity with credit risk concepts (PD/LGD/EAD, scorecards, thin-file lending) and with treasury, float and liquidity management.
- Awareness of regulatory expectations on model risk management, explainability and data protection in financial services.
Experience & Qualifications
- 12–18 years of overall experience in AI/ML, Data Science, Machine Learning Engineering, or related areas.
- Significant experience in building and deploying AI/ML solutions at scale, ideally in fintech, banking, payments or telecom.
- Demonstrated track record of taking models from PoC to production with measurable business impact.
- Bachelor's or Master's degree (PhD a plus) in Computer Science, Statistics, Mathematics, Data Science or a related field.
- Strong communication, stakeholder management, and technical leadership skills.
Success in the Role
Success in this role will be measured by the ability to:
- Identify and prioritise high-value AI/ML opportunities.
- Successfully convert AI/ML concepts and PoCs into production-ready product capabilities.
- Demonstrate measurable business impact, such as reduced fraud losses and false positives, improved forecast accuracy, and better credit decisions at controlled risk.
- Keep models reliable in production through continuous monitoring, drift management, retraining and governance.
- Accelerate AI adoption across Comviva's product portfolio.
- Build scalable, reliable, and commercially relevant AI solutions.
- Establish reusable AI/ML capabilities and best practices across engineering teams, and mentor the next generation of AI/ML talent.