[3:42 pm, 17/6/2026] Megha: About the role:
You are the engineer who turns models into reliable production systems running against live plant and enterprise data. You own the path from notebook to production — deployment, monitoring, retraining — including the demanding case of models that feed back into plant control systems where stability and latency are non-negotiable. Without this role, pilots never industrialise.
Key responsibilities
- Build and own CI/CD pipelines for ML; package, deploy and serve models in production (batch, real-time, and edge / on-prem as plant environments demand).
- Stand up and maintain MLOps infrastructure — experiment tracking, model registry, feature pipelines and automated retraining.
- Implement model monitoring (performance, data and concept drift, alerting) and ensure reliability and uptime in plant settings.
- Integrate models with operational systems — DCS / control systems for advisory and closed-loop use, plant historians — and with enterprise systems (SAP, Salesforce).
- Build robust data pipelines feeding the lakehouse from both OT and IT sources.
- Partner with data scientists to productionise their work, and with platform / security teams on OT/IT integration and cybersecurity.
- Champion engineering best practices: testing, versioning, reproducibility and documentation.
Required qualifications (must-have)
- Bachelor's or Master's in Computer Science, Engineering, or a related field.
- 3–6 years in ML engineering, MLOps, data engineering or software engineering with production ML exposure.
- Strong Python software-engineering fundamentals — OOP, type hints, automated testing (pytest), packaging, clean modular code, and version control.
- Solid working command of classical ML (scikit-learn pipelines, regularised regression,
tree-based ensembles, clustering) — enough to package, serve, optimise and monitor these models reliably in production.
- Hands-on with containerisation (Docker), orchestration (Kubernetes) and workflow tools (Airflow / Prefect / Dagster).
- Experience with MLOps tooling (MLflow, Kubeflow, SageMaker, Azure ML or similar) and model serving.
- Cloud experience (Azure, AWS or GCP) and a modern data stack (Spark / Databricks).
- CI/CD (Git-based pipelines) and familiarity with infrastructure-as-code.
Preferred (strong pluses)
- Experience deploying ML in industrial / manufacturing settings, including edge or on-prem deployment near plant equipment.
- Exposure to OT / IIoT integration — plant historians, OPC-UA, time-series databases, streaming (Kafka).
- Integration experience with SAP (ERP) and Salesforce (SFDC) data and APIs.
- Real-time / streaming ML and model optimisation for low latency.
- Awareness of OT cybersecurity considerations.
Technical skills
- Python software engineering: advanced, production-grade Python — OOP, type hints, automated testing (pytest), packaging & dependency management, profiling and performance; building model APIs with FastAPI / Flask; strong SQL and Git.
- Classical ML in production: working command of scikit-learn pipelines and classical models (regularised regression, tree-based ensembles — XGBoost / LightGBM, clustering); model serialisation (joblib / pickle / ONNX); batch and real-time inference.
- Data engineering: pandas and PySpark for batch processing; streaming with Kafka; data validation (pydantic, Great Expectations); feature pipelines / feature stores.
- MLOps & infrastructure: Docker, Kubernetes, MLflow, Airflow / Prefect, CI/CD and infrastructure-as-code; model registries and serving frameworks.
Platform & tooling
Cloud (Azure / AWS / GCP); Databricks / Spark; time-series databases and historian / OPC-UA connectors.
Pay: ₹398,175.91 - ₹1,652,806.44 per year
Work Location: In person
