Overview

We’re looking for an AI Engineer to build LLM-based systems that extract structured knowledge from documents and enable intelligent reasoning over complex systems. You’ll work at the intersection of document understanding, knowledge graphs, and RAG systems — turning unstructured knowledge (manuals, technical documents, domain-specific data) into queryable, actionable intelligence.

Responsibilities

• Design and build LLM-based pipelines for parsing and structuring information from technical documents and diagrams
• Build and maintain knowledge bases using vector and graph databases
• Build proprietary agentic systems for production deployment
• Integrate LLM and CV models into end-to-end extraction pipelines
• Develop graph-based RAG systems for reasoning over structured knowledge
• Create evaluation metrics and benchmarks to measure pipeline and RAG quality
• Research, prototype, and deploy model distillation approaches for production systems
• Iterate on prompt engineering, retrieval strategies, and pipeline orchestration
• Collaborate with research and software engineering teams

Requirements

• Hands-on experience building LLM-based applications (multi-step pipelines, not just API calls)
• Experience with RAG systems: embedding, retrieval, reranking, generation
• Familiarity with vector databases (Pinecone, Weaviate, Qdrant) and/or graph databases (Neo4j)
• Strong Python and ML tooling (PyTorch, HuggingFace, LangChain/LlamaIndex or similar)
• Experience with LLM output evaluation — metrics, benchmarks, human-in-the-loop
• Solid software engineering fundamentals — code quality, testing, version control

Nice to Have

• Foundation model fine-tuning (SFT, QLoRA, RLHF)
• Experience with document parsing / OCR pipelines (PDF extraction, diagram understanding)
• Knowledge graphs and ontologies
• Experience with LangChain / LangGraph
• Integrating CV models into multi-modal pipelines
• Background in industrial, scientific, or regulated domains
• Model distillation and compression techniques