About this role
AuxoAI seeks a Senior Applied Scientist to design and deploy structured knowledge systems that support reliable, schema-grounded AI and agent reasoning. The role combines large language models, knowledge graphs, semantic architectures and hybrid retrieval. The posting lists Bangalore, Hyderabad, Mumbai and Gurgaon/Gurugram as locations and indicates that the role may allow working from home.
Responsibilities
- Design schema-guided extraction using zero-shot and few-shot structured prompting, JSON-schema or grammar-constrained decoding, and function-calling or tool-driven techniques. Build multi-stage pipelines for nested entities, hierarchical structures and cross-document relationships.
- Develop ontology-driven systems with tools such as LinkML, OWL and SHACL, representing knowledge as RDF triples or labeled property graphs. Design entity resolution using blocking, embedding similarity and rules; develop ontology alignment and, where appropriate, graph embeddings such as Node2Vec or TransE.
- Build hybrid retrieval that combines dense and sparse retrieval with graph traversal, path ranking and neighborhood expansion. Validate schema conformance and semantic consistency to reduce invalid or hallucinated structured outputs.
- Integrate structured knowledge into GraphRAG, agent planning and tool-selection workflows. Deliver production systems with targets for latency, scalability, reliability and data integrity.
Requirements
The posting specifies 5–10 years of experience, including 5+ years building production AI or machine learning systems, and lists a bachelor’s degree in its structured job data. Candidates need experience with knowledge graphs or ontology-driven architectures; structured extraction using techniques such as constrained decoding, JSON-schema enforcement or AST-style parsing; entity resolution beyond simple embedding similarity; SPARQL or Cypher and graph-query optimization; RDF, OWL or property-graph models; and strong Python engineering focused on validation, schema integrity and reliability. Experience designing hybrid symbolic and neural AI systems is also required.
Preferred
Experience with PageRank, community detection or shortest-path algorithms; GraphRAG; compositional semantic extraction; reasoning engines or rule-based inference; and evaluation of structural extraction accuracy and consistency. Candidates able to join immediately or within two weeks are being prioritized.