New opportunity

AI Engineer (Energy)

John Cockerill · Mumbai, India

About this role

John Cockerill Energy seeks an AI Engineer to turn approved use cases into production-grade AI products across its Cooling, Energy Solutions and Energy Transition businesses. Reporting to the AI Product Manager, the engineer works within corporate AI governance, cybersecurity and technology standards as the organization moves from experimental agents to governed, scalable products.

Responsibilities

Design AI applications, copilots and multi-agent systems; translate business requirements into solutions and develop reusable components. Implement orchestration, tool calling, memory, guardrails, monitoring and human-in-the-loop controls.
Create and evaluate prompts and models; build retrieval-augmented generation (RAG), vector search and semantic retrieval solutions. Optimize accuracy, safety, latency, cost and business value.
Integrate AI products with enterprise systems, including SAP S/4HANA, JD Edwards, Planview, SharePoint, Teams, Microsoft 365 and engineering platforms. Build APIs and data pipelines while protecting enterprise data.
Deploy and operate solutions on approved Azure environments using CI/CD, infrastructure automation, logging and alerting. Manage AI product testing, validation, deployment, monitoring, version control and retirement; monitor token consumption and cloud costs.
Conduct quality evaluations, document limitations and mitigations, maintain architecture and support documentation, and assist with pilots, proofs of concept, production rollouts and troubleshooting. Follow Group AI Policy, Energy AI Governance SOP, cybersecurity and privacy requirements, and EU AI Act requirements.

Qualifications

Bachelor’s or master’s degree in engineering, computer science, data science, AI, software engineering, IT or a related discipline. At least 3 years in software, data or AI engineering or cloud solution development, including at least 2 years delivering AI/ML, generative AI, Copilot or advanced analytics solutions.
Experience deploying enterprise solutions on Microsoft Azure and working with global, cross-functional teams. Relevant technologies include Azure AI Foundry, Azure OpenAI, Copilot Studio, Microsoft Graph, Power Platform, Python, REST APIs, GitHub and CI/CD. Strong analytical, communication and problem-solving skills are required. Experience in energy, EPC, industrial manufacturing, engineering or capital goods is preferred.

Work arrangement: Based in the Navi Mumbai office, with up to 2 work-from-home days per week subject to discussion with the reporting manager. Periodic travel to Europe is required.

For full details about this role, click “Apply on company site” to view the employer’s original posting.

Skills for this role

Generative AILarge Language ModelsAgentic AIMulti-agent systemsMicrosoft Copilot StudioAzure AI FoundryAzure OpenAIMicrosoft GraphMicrosoft 365 CopilotPower PlatformAzure FunctionsAzure Logic AppsPrompt engineeringRetrieval-Augmented GenerationVector databasesEmbeddingsSemantic searchAI evaluationPythonREST APIsJSONGitHubCI/CDMicrosoft AzureInfrastructure as codeMLOpsLLMOpsSAP S/4HANAJD EdwardsPlanview Enterprise One

Checking your profile…

Preparation suggestions, not the employer’s actual interview questions. Always check the original posting for current requirements.

Understand the role

What matters most

Turn approved use cases into products
Under the AI Product Manager, translate user stories across Cooling, Energy Solutions and Energy Transition into production-grade AI applications, copilots and reusable components.
Engineer and evaluate grounded AI
Build agents with tool calling, memory, guardrails and human-in-the-loop controls; develop RAG and semantic retrieval; test outputs for accuracy, relevance, safety, latency, cost and business value.
Integrate enterprise systems securely
Connect AI products to systems such as SAP S/4HANA, JD Edwards, Planview, Microsoft 365 and engineering platforms through APIs and data pipelines while protecting enterprise data.
Operate AI across its lifecycle
Deploy on approved Azure environments using CI/CD, infrastructure automation and observability. Manage validation, monitoring, version control and retirement, including token and cloud-cost tracking.
Make solutions governable and maintainable
Follow Group AI Policy, Energy AI Governance SOP, cybersecurity, privacy and EU AI Act requirements. Document limitations, architecture and support procedures, and assist with pilots, rollouts and troubleshooting.

Likely challenges to think through

Agent autonomy versus control
A likely design challenge is allowing agents to complete useful workflows without giving tool calls excessive authority. Prepare to explain where permissions, guardrails, monitoring and human approval would sit in an agent workflow.
Answer quality versus operating cost
Grounded answers may require more retrieval and model work, increasing latency and token spend. Prepare an evaluation approach that compares answer quality and safety with latency, cost and business value.
Reuse versus system-specific integration
Shared components must still accommodate different enterprise systems and data-access rules. Prepare to describe how you would separate reusable AI services from integration adapters, then test and document the result for production support.

Practice your answers

Technical practice

How would you turn an approved AI use case into a production-ready solution?

What to coverThe role translates user stories into governed, scalable AI products. Suggested preparation: explain how you would clarify requirements, choose an architecture, identify reusable components, and define validation and operational readiness.

How would you design a multi-agent workflow with tool calling?

What to coverThe posting names orchestration, agent collaboration, memory, guardrails and human-in-the-loop controls. Suggested preparation: discuss when multiple agents are warranted, how tools and state are controlled, and where a person should approve consequential actions.

How would you ground an assistant in trusted enterprise content?

What to coverThe role includes RAG, embeddings, vector search, semantic retrieval, metadata and taxonomies. Suggested preparation: explain retrieval and access-control choices, then describe how you would check whether retrieved evidence actually supports an answer.

How would you evaluate and improve prompts and model outputs?

What to coverThe posting calls for documented prompts and evaluation of accuracy, relevance, safety, latency, cost and business value. Suggested preparation: describe representative test cases, failure analysis, and the trade-offs you would assess before changing a prompt or model.

How would you govern a copilot connected to Microsoft 365?

What to coverCopilot Studio, Microsoft Graph, Microsoft 365 permissions and collaboration data are named in the posting. Suggested preparation: explain how you would respect existing permissions, constrain available actions and test for unintended disclosure.

How would you integrate an AI product with enterprise systems?

What to coverThe role names SAP S/4HANA, JD Edwards, Planview and engineering platforms, along with APIs and data pipelines. Suggested preparation: discuss interface contracts, data quality, error handling and protection of enterprise data without assuming a particular system integration is assigned.

How would you deploy an AI application into an approved Azure environment?

What to coverThe posting calls for CI/CD, infrastructure automation, logging and alerting on approved cloud environments. Suggested preparation: walk through repeatable deployment, environment configuration, release checks and the signals needed to operate the service.

How would you manage an AI product throughout its lifecycle?

What to coverThe stated lifecycle spans development, testing, validation, deployment, monitoring and retirement, with version control and model governance. Suggested preparation: explain what you would version, how you would validate a change, and how you would handle rollback or retirement.

How would you investigate poor production performance or rising costs?

What to coverThe role includes production troubleshooting, usage analysis, token-consumption monitoring and cloud-cost monitoring. Suggested preparation: describe how you would distinguish retrieval, model, integration and operational failures while weighing quality, latency and cost.

How would you address security and compliance during AI solution design?

What to coverThe posting cites Group AI Policy, the Energy AI Governance SOP, cybersecurity, privacy and EU AI Act requirements, plus risk assessments and architecture reviews. Suggested preparation: explain how you would identify risks, document limitations and mitigations, and seek review rather than assume an unapproved design is compliant.

Behavioural practice

Tell me about a time you translated a business need into a technical solution.

What to coverThe role requires turning approved requirements into working AI solutions. Suggested preparation: use a real situation, the questions and decisions you personally handled, and the resulting outcome or lesson.

Describe a time you worked through a disagreement with cross-functional stakeholders.

What to coverThe posting emphasizes communication with technical and non-technical stakeholders in a global matrix organization. Suggested preparation: give the situation, your actions to clarify trade-offs and reach a decision, and the result.

Tell me about a time you identified a security, privacy or governance concern.

What to coverThe role participates in risk assessments and technical governance. Suggested preparation: describe the situation, how you raised and addressed the concern within your authority, and what happened.

Describe a production issue you took ownership of resolving.

What to coverThe posting includes rollout support, troubleshooting, monitoring and continuous improvement. Suggested preparation: give the situation, your diagnostic and communication actions, and the operational result or learning.

Tell me about a time user feedback changed your technical approach.

What to coverThe role calls for analyzing usage and feedback to recommend improvements that support business outcomes. Suggested preparation: explain the situation, how you evaluated and acted on the feedback, and the observed result.

Sharpen your knowledge

Topics to prepare

Governed agent architecture
Posting: Design copilots and multi-agent systems with tool calling, memory, guardrails, monitoring and human-in-the-loop controls. Revision exercise (suggestion): Sketch an agent workflow for a hypothetical approved use case, marking where tools can act, where a person must approve, and how failures are handled.
Enterprise retrieval and knowledge design
Posting: Ground AI outputs in trusted content using RAG, embeddings, vector search, semantic search, metadata and taxonomies. Revision exercise (suggestion): Diagram how documents would be indexed and retrieved, then explain how metadata and source permissions would affect the answer.
AI evaluation and trade-offs
Posting: Assess accuracy, relevance, safety, latency, cost and business value; document limitations and mitigations. Revision exercise (suggestion): Define a small test set for a hypothetical assistant, specify measures for each relevant dimension, and explain how you would compare two prompt versions.
Security, governance and lifecycle
Posting: Follow Group AI Policy, Energy AI Governance SOP, cybersecurity, privacy and EU AI Act requirements while managing products through validation, monitoring and retirement. Revision exercise (suggestion): Map these stated review areas onto a release-to-retirement checklist without assuming the contents of internal policies.
Enterprise integration boundaries
Posting: Connect AI products to systems including SAP S/4HANA, JD Edwards, Planview and Microsoft 365 while protecting enterprise data. Revision exercise (suggestion): Draw an API-based integration boundary and identify authentication, data-access and failure-handling questions to resolve before implementation.

Skills to revise

Build and test a retrieval workflow
Posting: Develop RAG, vector search and semantic retrieval solutions. Practice suggestion: Using sample documents, assemble a small retrieval-to-answer workflow and test whether answers cite relevant retrieved content and decline unsupported requests.
Implement controlled tool calling
Posting: Build agent interfaces, workflows, tool calling and human-in-the-loop controls. Practice suggestion: Prototype a read-only tool call with input validation, an approval step before any proposed change, and logs showing the decision path.
Create an Azure deployment plan
Posting: Deploy enterprise solutions on approved Azure environments using CI/CD, infrastructure automation, logging and alerting. Practice suggestion: Draft a deployment pipeline and infrastructure-as-code outline, identifying validation gates, rollback steps and operational alerts; do not assume access to the employer's environment.
Develop a secure integration interface
Posting: Use Python, REST APIs and JSON to build AI services and enterprise integrations. Practice suggestion: Write a small Python API against mock enterprise data, with schema validation, access checks and tests for malformed requests.
Inspect AI product operations
Posting: Monitor model performance, token consumption, cloud costs and production issues. Practice suggestion: Create a sample operational dashboard specification covering quality results, latency, failures, token use and cost, then describe which signals would trigger investigation.

Your preparation checklist

Map the role to your evidence
Prepare a one-page evidence map for the stated experience, Azure deployment, AI delivery and cross-functional requirements. Mark any sector experience separately; energy or related industry experience is preferred, not required.
Prepare an AI product walkthrough
Choose an actual project, if available, and outline the business requirement, your contribution, architecture, rollout and measured results. The role focuses on turning approved use cases into governed, production-grade products; do not invent missing outcomes.
Practice an agent design exercise
Sketch a hypothetical assistant for an approved Energy use case, showing tool calling, memory, guardrails, monitoring and human approval points. Label it as a preparation exercise, not a company project.
Build a RAG evaluation plan
Draft a small test set and scorecard for retrieval relevance, answer accuracy, safety, latency and cost. Explain how you would document limitations and mitigations, as the posting calls for model evaluation and quality validation.
Diagram a secure integration
Show how a proposed AI service could access permitted content from SharePoint or Microsoft 365 and connect through APIs to an enterprise system. Identify permission checks, data protection boundaries and failure handling; the posting names these integration areas but does not specify an implementation.
Outline the production lifecycle
Prepare an Azure deployment diagram covering version control, CI/CD, testing, logging, alerting, token and cloud-cost monitoring, and retirement. Note where architecture review, AI governance, cybersecurity, privacy and EU AI Act checks would be addressed.
Write employer questions
Bring questions about how approved use cases are prioritized, which Azure services and enterprise integrations are approved, how evaluation and human oversight are handled, and how the AI Product Manager, business lines and Group IT coordinate rollouts.

Availability last checked: 5 Nov 2026. Status can change after a check; confirm on the employer’s website. Report an issue · Our checking policy

KEEP EXPLORING

Similar AI jobs

Related skills and specializations in India. Matched to this role, not your profile.

Explore more jobs

Finding similar opportunities…