AI Context Architect

Role Name: AI Context Architect
Experience : 12+ Years

Relevant Experience : 2-3+ Years

Location : Chennai or Bengaluru

1. Role Summary

The AI Context Architect makes enterprise knowledge trustworthy and usable by AI systems. The role designs the semantic layer (ontologies, taxonomies and semantic models), architects retrieval on top of it (GraphRAG and hybrid graph-vector patterns across graph databases, vector stores, RDBMS and NoSQL), governs it (metadata, lineage, entity resolution, access and explainability), and exposes it as trusted context to agents and applications through APIs and MCP.

The role works as a hands-on, AI-native engineer-architect using agent harnesses in daily delivery, and sets reference patterns that Solutions and Advisory teams reuse with clients.

2. Key Responsibilities

A. Semantic & Knowledge Modeling

• Define enterprise ontologies, taxonomies, and semantic models that represent business entities, relationships, and rules in a form both people and machines can use.
• Establish modeling standards, naming conventions, and versioning/change processes for the semantic layer.
• Translate domain and business concepts into graph and relational models with domain experts.

B. Retrieval & Context Architecture

• Architect GraphRAG and hybrid graph-vector retrieval solutions across Neo4j, Amazon Neptune, vector stores, RDBMS, and NoSQL.
• Design chunking, embedding, indexing, ranking, and query-planning strategies; select the right retrieval pattern per use case (vector, graph traversal, SQL, hybrid).
• Design context assembly for agents: what to retrieve, how to rank, how to cite, and how to fit within context and cost limits.

C. Governance, Trust & Explainability

• Govern metadata, data lineage, and provenance for all context served to AI systems.
• Design and implement entity resolution and master-data alignment across sources.
• Enforce access control (row/entity/attribute-level, policy-aware retrieval) so AI only sees what the requester is entitled to see.
• Build explainability into retrieval: traceable sources, reasoning paths, and confidence indicators.

D. Exposure & Integration

• Expose trusted context as secure, versioned APIs and MCP servers consumable by agents, copilots, and applications.
• Apply cloud-native, microservices, and API design principles for scalable, resilient context services.
• Integrate with enterprise data platforms, catalogs, and identity/security systems.

E. Quality, Operations & Cost

• Define evaluation frameworks for retrieval and context quality (relevance, groundedness, completeness, latency, regression testing).
• Instrument context services for observability (tracing, quality metrics, usage, drift).
• Manage cost and performance trade-offs (index sizes, model calls, query patterns, caching).

F. Engineering Practice & CoE Leadership

• Build working reference implementations and accelerators using AI-native engineering practices and agent harnesses.
• Apply DevSecOps practices: CI/CD, infrastructure as code, automated testing, secure-by-design.
• Publish reference architectures, patterns, and playbooks; mentor engineers and support Solutions and Advisory teams in client engagements.
• Evaluate emerging tools, standards, and models, and advise on adoption.

3. Required Skills & Qualifications

Education & Experience

• Bachelor's or Master's degree in Computer Science, Information Systems, Data Science, or an equivalent field, or equivalent practical experience.
• 10+ years in software, data, or solution architecture, with at least 3 years in knowledge graphs, semantic modeling, search/retrieval, or applied AI/LLM systems.
• Track record of taking at least one graph- or retrieval-based AI solution into production.

Technical Skills (Must Have)

1. Semantic modeling: ontologies, taxonomies, semantic models, knowledge representation, property graph modeling.
2. Graph technologies: hands-on with Neo4j and/or Amazon Neptune; Cypher, Gremlin, or SPARQL.
3. Retrieval architecture: GraphRAG and hybrid graph-vector retrieval; embeddings, vector stores, re-ranking, query planning.
4. Polyglot data: RDBMS (SQL design and optimization) and NoSQL (document, key-value, wide-column) and how to combine them in retrieval.
5. Governance: metadata management, lineage, entity resolution, access control, and explainability for AI.
6. Context exposure: designing and building APIs and MCP servers for agent consumption.
7. AI-native engineering: practical use of AI in the development lifecycle; hands-on with at least one agent harness (Claude Code, GitHub Copilot, Codex, Kiro, or Cursor).
8. Modern application development: strong proficiency in at least one language such as Python, TypeScript/Node, or Java; clean architecture, testing, code review.
9. Cloud-native engineering: AWS/Azure/GCP services, containers, microservices, API design, event-driven patterns.
10. DevSecOps: CI/CD, IaC, secrets management, secure SDLC, dependency and supply-chain hygiene.
11. Enterprise AI hygiene: security, governance, evaluation, observability, and cost awareness for LLM-based systems.

Behavioral & Professional Skills

• Structured, first-principles thinking; able to turn ambiguous business domains into clean models.
• Clear communication with executives, domain experts, and engineers; strong written architecture documentation.
• Consultative mindset; balances elegance with delivery pragmatism.
• Ownership, mentoring, and a collaborative approach in a CoE setting.

4. Other Skills (Good to Have)

Role-specific

• Semantic web standards: RDF, OWL, SHACL, SKOS, JSON-LD.
• Graph algorithms and graph data science (community detection, centrality, embeddings).
• LLM-assisted knowledge graph construction (entity/relation extraction, ontology induction).
• Data catalog and governance platforms (e.g., Collibra, Alation, DataHub, OpenMetadata).
• Evaluation tooling for RAG/agents (e.g., RAGAS, TruLens, LangSmith, custom harnesses).
• Orchestration and agent frameworks (LangGraph, LlamaIndex, Semantic Kernel, custom agents).
• Search technologies (OpenSearch/Elasticsearch, BM25 + vector hybrid search).
• Streaming and data pipelines (Kafka, Spark, dbt, Airflow).
• Domain exposure in BFSI, healthcare/life sciences, manufacturing, retail, or public sector.

Common competencies (mandatory across all roles)

• Practical AI-native engineering and strong modern application development.
• Cloud-native, API, microservices, and DevSecOps grounding.
• Hands-on experience with at least one agent harness (Claude Code, GitHub Copilot, Codex, Kiro, Cursor).
• Enterprise security, governance, evaluation, observability, and cost awareness.

Certifications (preferred, not mandatory)

• Cloud architect certifications (AWS, Azure or GCP).
• Neo4j Certified Professional / Graph Data Science certification.
• TOGAF, DAMA CDMP, or equivalent architecture/data management certification.

4. Success Measures (Indicative)

• Reference architecture for GraphRAG / hybrid retrieval and an enterprise semantic model approved and reusable across engagements.
• Context services exposed via API/MCP with lineage, access control, and explainability in place.
• Retrieval quality tracked with a defined evaluation suite and demonstrable improvement over baseline.
• Accelerators, patterns, and playbooks adopted by Solutions and Advisory teams.
• Measurable contribution to client outcomes: accuracy, trust, time-to-deploy, and cost per query.