AI Engineer - Senior

MXI works with complex construction records to support factual analysis and litigation workflows. We’re hiring an AI Engineer to build systems that make large collections of documents searchable, connect related evidence, and produce accurate answers with traceable sources., * Build and maintain RAG pipelines covering document ingestion, parsing, chunking, embeddings, indexing, retrieval, reranking, and answer generation.

• Work with PDFs, emails, spreadsheets, reports, and other structured and unstructured records.
• Improve retrieval using keyword search, vector search, metadata filters, and hybrid approaches.
• Preserve document relationships, versions, and source references so users can verify answers against the original records.
• Develop evaluation datasets and measure retrieval quality, answer accuracy, citation correctness, latency, and cost.
• Investigate failure cases, reduce unsupported answers, and build appropriate handling for incomplete or conflicting evidence.
• Integrate AI capabilities into backend services and user workflows through APIs.
• Maintain reliable processing pipelines, monitoring, access controls, and secure handling of confidential documents., Users can find relevant records, receive answers supported by accurate citations, and understand when the available evidence is insufficient. The system handles new documents reliably, and improvements are demonstrated through repeatable evaluations.

We’re looking for someone with hands-on experience in retrieval-augmented generation (RAG), strong Python skills, and the ability to take AI features from an initial prototype into reliable production use., * Practical experience building RAG or document search systems, with examples you can explain in technical detail.

• Strong Python engineering skills, including APIs, data processing, testing, and debugging.
• A solid understanding of embeddings, vector databases, keyword search, chunking strategies, and reranking.
• Experience integrating LLM APIs and managing context, structured outputs, and model responses.
• Working knowledge of SQL, Git, cloud infrastructure, and application deployment.
• The ability to evaluate system quality systematically and diagnose whether a problem comes from the data, retrieval, or generation.
• Clear communication, sound judgment, and the ability to own a problem through implementation and verification.

Helpful experience

• Search and analysis across large document collections.
• Tools such as PostgreSQL/pgvector, Elasticsearch/OpenSearch, Qdrant, or comparable systems.
• Frameworks such as LangChain or LlamaIndex, alongside an understanding of the underlying components.
• Document deduplication, entity extraction, metadata enrichment, and version tracking.
• Construction, legal technology, enterprise search, or other fields where accuracy and source traceability matter.