AI Architect

Aaura Softwares is hiring a Lead AI Architect to own knowledge graph architecture across client engagements, from ontology design and technology selection through production delivery. The role also builds reusable graph and GraphRAG practices and supports technical pre-sales.

Responsibilities:

• Own the knowledge graph architecture end to end on assigned engagements — conceptual model, ontology or schema, graph data model, ingestion and enrichment pipelines, query and serving layer
• Design entity and relationship models from messy enterprise reality: multiple source systems, inconsistent identifiers, partial records, and business definitions that differ by department
• Define the entity resolution strategy — blocking, matching, survivorship, and the human-in-the-loop path for the cases the algorithm should not decide alone
• Architect GraphRAG and graph-grounded retrieval — hybrid vector plus graph traversal, subgraph retrieval, context assembly, and prompt grounding strategies that keep answers traceable to graph facts
• Make and defend build-versus-buy calls across graph databases, vector stores, ontology tooling, and orchestration frameworks, with a written rationale a client architect can challenge
• Set the non-functional bar — query latency, graph scale, incremental update patterns, lineage, access control at node and relationship level, and how the graph is versioned as the ontology evolves
• Lead technical discovery: work with client SMEs to extract the domain model that exists in their heads and nowhere in their systems
• Build proofs of concept yourself. At this level we expect you to be able to open an editor and demonstrate the pattern, not only describe it
• Guide engineering teams through implementation — design reviews, Cypher and pipeline reviews, unblocking hard modelling questions
• Own the path from pilot to production: evaluation harnesses, monitoring, drift and quality checks on graph content, and clear operational ownership before go-live
• Set graph quality metrics and hold the delivery to them — coverage, resolution precision and recall, answer groundedness, and retrieval relevance
• Build reusable accelerators, reference architectures, and delivery playbooks for graph and GraphRAG engagements
• Support pre-sales: solution shaping, effort and cost estimation, architecture sections of proposals, and technical defence in client evaluations
• Mentor architects and senior engineers into graph modelling competence — this capability cannot rest on one person
• Represent the practice externally where useful: client architecture forums, partner sessions, conference talks, published reference material