Responsibilities
- Create and deploy generative AI and agent-based systems for interpreting complex documents, logical reasoning, and decision-making assistance
- Construct and refine retrieval-augmented generation workflows specialized for legal and regulatory texts, ensuring accurate and fact-based outputs
- Design reliable methods for ingesting and retrieving documents using context-aware segmentation, embedding models, metadata enhancement, and semantic indexing
- Implement systems to track references, citations, and traceability across AI-driven document processing pipelines
- Enhance retrieval ranking, semantic search accuracy, and response grounding to increase correctness and minimize false information generation
- Combine Knowledge Graph technologies (RDF/SPARQL) with large language model workflows to enable reasoning across structured and unstructured data
- Coordinate multi-stage AI processes using frameworks such as LangChain or LangGraph for agent-based logic
- Develop evaluation protocols for AI quality, including retrieval performance, hallucination detection, LLM-based judging, and metrics like RAGAS
- Design, train, and fine-tune custom named entity recognition and document comprehension models
- Ensure AI outputs are transparent, auditable, and compliant within highly regulated domains
- Support full lifecycle model development, including experimentation, version control, deployment preparation, and monitoring transition
