GenAI Engineer - Multimodal Document Intelligence

MaintainX; New York (NY); On-site; USD 152,000 - 271,000; Full time

Job summary

MaintainX, now part of Autodesk, seeks a GenAI-focused engineer to own the Generation half of Document Intelligence. You will turn multimodal primitives into schema-valid entities, ensuring high-quality outputs and maintaining strict quality controls across prompts, schemas, and evaluation pipelines.

You will build prompts, schemas, and eval rubrics, and manage model selection and budgeting to balance quality, cost, and latency.

Qualifications

• Strong applied GenAI craft with prompt engineering and structured outputs.
• Experience building eval datasets and rubrics, tracking quality.
• Shipped LLM features into production rather than notebooks.
• Comfort handling video, PDF, audio and image inputs converted to text.

Responsibilities

• Design and iterate prompt recipes (system, few-shot, context).
• Define per-entity schemas and validators for generated outputs.
• Build eval datasets and rubrics; monitor quality and regressions.
• Assemble multimodal context windows for model calls.
• Choose models and budget per recipe based on quality and cost.

Skills

Prompt engineering

Structured outputs

RAG retrieval patterns

Eval datasets

Rubrics

LLM production

Multimodal input

LLM observability

OCR & document understanding

Fine-tuning

Job description

MaintainX, now part of Autodesk, seeks a GenAI-focused engineer to own the Generation half of Document Intelligence. You will turn multimodal primitives into schema-valid entities, ensuring high-quality outputs and maintaining strict quality controls across prompts, schemas, and evaluation pipelines.

You will build prompts, schemas, and eval rubrics, and manage model selection and budgeting to balance quality, cost, and latency.