Senior AI Engineer | Remote | Jobs In India

What matters most

• Agentic semantic workflows

  The posting calls for agentic pipelines for semantic mapping, dimension mining, and ontology-driven reasoning, plus LLM services for schema understanding and metadata generation. Prepare to explain the design and evaluation of a comparable workflow.

• Enterprise ontology and graph reasoning

  The role builds RDF and OWL ontologies and SPARQL reasoning layers, with accuracy, traceability, and consistent interpretation emphasized. Prepare to discuss how definitions are modeled, queried, and validated; CIM alignment applies where relevant.

• Business-ready dimensional data

  The role connects semantic reasoning to Python services and batch pipelines, then develops Fabric dimension and fact pipelines and star schemas from heterogeneous sources. Prepare to trace how semantic definitions become usable data transformations.

• Production engineering standards

  The posting assigns reusable components and standards for quality, observability, security, and performance across cloud-based AI and data platforms. Prepare to explain operational design choices, not just prototypes.

• Technical leadership without people management

  The individual-contributor role includes reviewing code, mentoring engineers, making technical decisions, and working with architects, domain experts, and governance teams on definitions and change. Prepare examples of technical decisions reached across those groups.

Likely challenges to think through

• Balancing AI flexibility with semantic correctness

  Potential challenge, not a confirmed employer problem: LLM-generated mappings and metadata may be useful but hard to reconcile with traceable ontology reasoning. Prepare an approach for checking outputs against graph definitions and handling uncertain matches.

• Turning varied sources into stable dimensions

  Potential challenge, not a confirmed employer problem: heterogeneous inputs may use conflicting meanings or levels of detail, while star schemas need consistent facts and dimensions. Prepare to explain how you would resolve definitions with domain experts before encoding Fabric pipelines.

• Setting standards while delivering a greenfield platform

  Potential challenge, not a confirmed employer problem: ambiguous initiatives can pressure an individual contributor to ship quickly while establishing reusable, observable, secure components. Prepare to explain how you would sequence an initial implementation and the standards needed to operate it.

Topics to prepare

• RDF, OWL and enterprise ontology design

  The posting requires ontology and knowledge-graph knowledge, including upper ontologies; CIM alignment applies where relevant, and CIM experience is preferred. Revise by sketching an upper ontology and a domain extension for two sample source schemas, then explain which relationships OWL can infer.

• Agentic and LLM-based semantic workflows

  The role calls for agentic pipelines and LLM services for schema understanding, semantic mapping, dimension mining and metadata generation. Revise by diagramming how an agent proposes a mapping, checks it against an ontology and handles an ambiguous match.

• Semantic accuracy and traceability

  The posting emphasizes accurate, traceable and consistent semantic interpretation. Revise by defining review criteria for proposed schema alignments: supporting source fields, ontology terms, conflicting evidence and cases requiring domain-expert validation.

• ETL/ELT and dimensional modeling

  ETL/ELT, metadata-driven pipelines and star schemas are required concepts; the role produces dimensions and facts from heterogeneous sources. Revise by choosing a fact-table grain and explaining how two differently named source fields would map to a shared dimension.

• Production platform design and governance

  The role involves cloud-based AI and data platforms, engineering standards, and collaboration with architects, domain experts and governance teams. Revise by outlining a change to a semantic definition and how you would assess its downstream effects, quality, security, performance and observability.

Skills to revise

• Build an agentic mapping pipeline

  As practice for the stated pipeline responsibility, use Google ADK, LangChain agents or a similar framework named by the posting to route sample schema fields through mapping, ontology checks and an explicit unresolved-case outcome.

• Write graph queries and reasoning checks

  Create a small RDF/OWL graph and write SPARQL queries that retrieve proposed field-to-concept mappings and the relationships supporting them. This directly rehearses the posted knowledge-graph querying and reasoning work.

• Implement a Python semantic service

  Build a small Python service or batch step that accepts a schema description and returns proposed semantic labels with supporting evidence and uncertain cases. This is a practice exercise for the required Python, LLM-system and AI-assisted metadata work, not a specified employer project.

• Produce dimension and fact pipelines

  Using sample heterogeneous tables, draft a Microsoft Fabric Lakehouse pipeline with Spark or SQL transformations that yields a dimension, a fact table and checks on keys and fact grain. The posting explicitly names Fabric, Lakehouse, Spark, SQL and star-schema delivery.

• Review a production-oriented change

  Practice a code review of a hypothetical ontology or pipeline change: identify a semantic regression, a traceability gap and an operational check, then explain the decision to a domain expert. This reflects the posted review, mentoring and cross-team responsibilities.

• Map relevant experience to the role

  Prepare a one-page evidence map with specific examples of Python, LLM systems, agentic frameworks, RDF/OWL/SPARQL, data pipelines, and cloud operations. Mark gaps honestly; the posting lists these as required skills.

• Rehearse an end-to-end project walkthrough

  Select a real project, if available, and explain the problem, your decisions, implementation, results, and trade-offs. Emphasize any schema understanding, semantic mapping, or AI-assisted metadata work relevant to the advertised responsibilities.

• Sketch a semantic mapping design

  As a preparation exercise, diagram how an agentic workflow could map heterogeneous schemas to an RDF/OWL ontology, use SPARQL for checks, and retain traceable human-reviewed decisions. Explain how you would test accuracy and handle ambiguous mappings.

• Work through a star-schema example

  Draft a small example showing source data, dimension and fact definitions, and ETL/ELT steps. Be ready to discuss how it could run with Microsoft Fabric Lakehouse, Spark, SQL, and orchestration tools named in the posting.

• Prepare an operational design review

  Outline a Python service and batch pipeline connecting semantic reasoning to data transformations. Identify concrete approaches to quality, observability, security, performance, and change handling; distinguish approaches you have used from proposals.

• Collect technical leadership examples

  Prepare brief, specific accounts of code reviews, mentoring, technical decisions, or work with data architects and domain experts. The role is an individual contributor position with these collaboration responsibilities, not a stated people-management requirement.

• Write questions for the employer

  Ask which semantic models and data sources are in scope, how definitions are validated with domain and governance teams, what Azure or Fabric environment exists, and how success is measured for the client’s greenfield work. These are questions to clarify, not assumptions about the project.