Senior Applied Scientist - Semantics

About the Role

AuxoAI is actively seeking a Senior Applied Scientist to architect and implement sophisticated structured knowledge systems enabling dependable, schema-driven AI and agent reasoning processes. This position bridges the domains of advanced large language models, knowledge graphs, semantic frameworks, and hybrid retrieval mechanisms.

The successful candidate will be responsible for developing solutions that convert unstructured information into precise structured knowledge forms, enforce semantic constraints rigorously, and facilitate hybrid symbolic-neural reasoning within production-grade environments.

This role will focus on creating scalable semantic infrastructures that power cutting-edge AI use cases such as GraphRAG pipelines, structured data extraction, and complex agent reasoning workflows. You will address challenges where conventional systems fall short, innovating by integrating machine learning, knowledge graph technologies, semantic rules, and classical AI methods to deliver robust, reliable production systems.

The position is based in Mumbai, Bangalore, Hyderabad, or Gurgaon with a hybrid work model requiring three days per week in the office.

Key Responsibilities

• Design and implement schema-guided information extraction systems leveraging zero-shot/few-shot structured prompts, constrained decoding (using JSON schema enforcement or grammar-based methods), and function-calling or tool-assisted extraction.
• Develop recursive and multi-stage extraction pipelines capable of managing nested entities, hierarchical structures, and relationships spanning multiple documents.
• Create ontology-driven systems employing frameworks such as LinkML, OWL, SHACL, or comparable schema modeling tools, with knowledge representation via RDF triples or labeled property graphs.
• Construct and enhance entity resolution algorithms incorporating methods like blocking, embedding similarity, and rule-based matching.
• Develop ontology alignment strategies and graph embeddings (e.g., Node2Vec, TransE) where appropriate.
• Design hybrid retrieval architectures that integrate dense vector retrieval, sparse retrieval techniques, and graph traversal algorithms including BFS, DFS, path ranking, and neighborhood expansion.
• Implement validation systems to ensure schema compliance, identify semantic inconsistencies, and minimize hallucinations or invalid structured outputs.
• Integrate structured knowledge components into GraphRAG pipelines, agent planning models, and tool-selection workflows.
• Deliver production-ready semantic systems optimized for latency, scalability, reliability, and data integrity.

Essential Qualifications and Experience

• Minimum of five years developing production-level AI or machine learning systems.
• Extensive experience in knowledge graph construction and ontology-driven architectural design.
• Proficient in structured extraction methodologies such as grammar-constrained decoding, JSON schema validation, and abstract syntax tree parsing.
• Skilled in developing advanced entity resolution beyond purely embedding similarity approaches.
• Proficient in graph query languages like SPARQL and Cypher, with expertise in tuning queries for performance.
• Familiar with RDF, OWL, and property graph data models and semantic architecture best practices.
• Strong software engineering skills in Python emphasizing data validation, schema correctness, and systems reliability.
• Experience designing AI systems that combine symbolic and neural techniques effectively.

Preferred Additional Experience

• Knowledge of graph algorithms such as PageRank, community detection, or shortest-path for reasoning chain development.
• Hands-on experience building graph-enhanced retrieval systems like GraphRAG.
• Designing compositional semantic extraction pipelines.
• Implementation of reasoning engines or rule-based inference systems.
• Expertise in benchmarking and validating the accuracy and consistency of structured extraction.

Skills and Technologies

• Machine Learning
• Semantics
• SPARQL query language
• JSON and Ontology engineering
• Knowledge Graphs
• Python programming
• Schema and data validation
• Graph traversal algorithms