Draft n Craft - Senior AI Engineer - RAG/LLM

Overview

• Design and build end-to-end document intelligence pipelines using OCR + LLMs.
• Parse, segment, and structure complex legal documents into schema-driven JSON outputs.
• Implement layout-aware extraction, table handling, and multi-page document processing.
• Apply prompt engineering, function calling, and structured output constraints.
• Build RAG-based systems (Retrieval-Augmented Generation) to ground LLM responses using internal legal knowledge bases.
• Fine-tune Large Language Models (LLaMA, GPT, and other open-source or commercial LLMs) for legal domainspecific tasks.
• Handle OCR noise, incomplete documents, inconsistent layouts, and edge cases.
  
  

• Integrate OCR engines such as AWS Textract, Azure Form Recognizer, or Tesseract.
• Design post-OCR cleanup, normalization, and rule-based extraction pipelines.
• Implement validation logic to ensure accuracy, consistency, and legal usability.
  
  

• Build scalable backend services using Python.
• Optimize latency, throughput, and cost (OCR usage, LLM token consumption).
• Implement logging, retries, failure handling, and confidence checks.
  
  

• Deploy AI systems on AWS / Azure / GCP.
• Collaborate with DevOps teams on containerization, monitoring, and secure access.
• Ensure solutions meet data security, privacy, and standard compliance requirements (e.g., HIPAA, where applicable).
  
  

• Work closely with legal professionals to translate domain requirements into technical solutions.
• Provide technical guidance and mentoring to junior engineers.
• Maintain clear technical documentation and architectural decision records.
  
  

(ref:hirist.tech)