ML Operations Engineer

Positions: 2 full-time roles

Location: India

Experience: 8–11 years overall

Role focus: Traditional data science and machine learning models (not LLMs or generative AI)

Job Description

Position Summary

Operate traditional data science and machine learning models in production as part of the India ML operations team. Responsibilities include executing deployments and retraining, monitoring for drift, maintaining lineage, and helping keep predictive models—such as churn, fraud, and network forecasting models—accurate and safe. This is a hands-on execution role that follows the standards and approvals set by the ML Operations Lead. The role does not cover LLMs or generative AI agents.

Role Details

• Job level/grade: P3 or P4
• Division: AI & Agent Operations (Day 2), AI Organisation
• Reports to: ML Operations Lead (India)

Key Responsibilities

• Execute model deployments and retraining through approved pipelines, including packaging, validation runs, and staged promotion.
• Monitor data, feature, and concept drift; prepare reports and flag threshold breaches.
• Validate models against the incumbent using backtesting or champion-challenger methods before promotion.
• Run optimization algorithms to select the best intervention candidate based on model predictions.
• Maintain current model lineage records, including version, data, baseline, and last-known-good information.
• Execute rollbacks to a previous model or safe baseline when instructed.
• Prepare model-performance, cost, and quota reports for the portfolio.
• Promptly raise data-quality issues with supporting evidence to the ML Operations Lead and Data Engineering team.

Typical Challenges

• Managing business-driven changes to model-generated intervention results.
• Handling model retraining and rerun cycles in response to business updates.
• Keeping models accurate as data and conditions change, and detecting drift early.
• Working with dependencies on upstream data quality.

Decision-Making Authority

Within own authority:

• Execute deployments, retraining, and monitoring within approved plans.
• Maintain lineage records and reporting.
• Escalate data issues promptly with supporting evidence.

Requires higher approval:

• Model-promotion approval from the ML Operations Lead.
• Changes to retraining strategy or thresholds.
• Rollback decisions outside agreed procedures.

Qualifications and Experience

Education: Degree in Computer Science, Data Science, Engineering, or a related field.

Experience: The posting lists 8–11 years overall; the qualifications section specifies 4–7 years in MLOps, DevOps, data engineering, or ML engineering. It also requests hands-on experience with churn predictive models, CI/CD pipelines, and production model deployment, as well as experience with drift monitoring or model-performance reporting.

Technical skills:

• Python and CI/CD tooling
• Model deployment, registries, and lineage concepts
• Data processing and reporting
• Model validation and drift monitoring
• Champion-challenger methods and backtesting
• Cloud ML services

Professional skills:

• Precision, reliability, and process discipline
• Clear status communication across a cross-border team
• Collaboration with data teams

Good to have: ML or cloud certifications; experience in a telco or large-scale data environment; and awareness of telco use cases such as churn, fraud, and forecasting.