Senior AI Context Engineer

Description

Our customer is seeking a Senior AI Context Engineer to help build and optimize the data foundations that power enterprise AI solutions. This role sits at the intersection of software engineering, applied AI, and data architecture, focusing on transforming structured and unstructured content, including policy documents, benefit materials, and clinical records, into high-quality context for AI applications.

The ideal candidate will design and enhance Retrieval-Augmented Generation (RAG) pipelines, vector search architectures, semantic retrieval frameworks, and context engineering processes that enable AI agents to deliver accurate, relevant, and compliant responses. Working closely with AI architects, developers, and business stakeholders, this individual will play a key role in advancing their AI capabilities while ensuring HIPAA-compliant access to enterprise knowledge and healthcare data.

Knowledge, Skills, And Abilities

Experience: 5+ years of experience in backend software engineering, applied machine learning/search, or data-centric application development.

GenAI & RAG Focus: 12 years of hands-on experience specifically optimizing AI and RAG architectures, prompt context strategies, and vector retrieval pipelines for LLM applications.

Programming & Tooling: Advanced proficiency in Python and SQL, alongside familiarity with modern orchestration tools (e.g., Airflow, Prefect, Temporal).

Unstructured Content Parsing: Deep experience using parsing frameworks (e.g., LlamaIndex, Unstructured.io, LangChain document loaders) to process complex layouts, tables, and unstructured documents.

Vector & Search Engines: Hands-on experience working with vector databases, embeddings, and semantic search platforms (e.g., Pinecone, pgvector, Weaviate, Qdrant, Azure AI Search).

Evaluation Frameworks: Familiarity with RAG and LLM context evaluation frameworks (e.g., Ragas, TruLens, Arize Phoenix) to measure retrieval recall, precision, and groundedness.

Data Security & Privacy: Practical experience handling sensitive healthcare data (PHI/PII) within high-compliance software environments.

Duties And Responsibilities

Context Engineering & Retrieval Optimization

Extraction & Chunking: Build and refine advanced structured and unstructured information parsing, layouts processing, and chunking workflows (converting PDFs, clinical notes, data, and policy docs) into high-quality contextual units for LLMs in partnership with cross-functional teams.

Semantic Search & Reranking: Implement hybrid search mechanics, metadata routing, and reranking logic to drastically improve retrieval precision and minimize model hallucinations.

Agentic Context Services: Design context payload specifications and metadata structures that feed into Model Context Protocol (MCP) servers and LLM orchestration tools built by application developers.

Vector Indexing & Retrieval Architecture

Index Design & Optimization: Recommends and implements vector indexing strategies, embedding schemes, and semantic query designs inside enterprise-provisioned vector stores (e.g., Pinecone, pgvector, Azure AI Search) to ensure high-performance, low-latency retrieval.

Vector Metadata: Design and manage sophisticated metadata tagging schemes to enable precise filtering, hybrid search, and domain-specific context retrieval.

Retrieval Evaluation & Groundedness: Establish automated evaluation frameworks to continuously monitor retrieval relevance, context quality, groundedness scores, and embedding drift over time.

Context Security, Compliance & Governance

Healthcare Context Compliance: Ensure all document processing and contextual payloads strictly adhere to HIPAA and HITRUST standards, implementing automated masking and tokenization for Protected Health Information (PHI).

Context-Level Access Control: Ensure role-based access control (RBAC) rules within vector metadata, ensuring AI search queries only return contextual snippets that the active user is authorized to see.

Auditing & Hand-off Lineage: Maintain context tracking and audit logs for prompt payloads, establishing clean data hand-off specifications when transitioning validated innovation prototypes to enterprise Data & Analytics or IT teams.

Collaborative Execution

Reference Pattern Alignment: Build upon the reference architectures, CI/CD templates, and “golden paths”…