Description

Build and optimize the on-premise AI platform for document intelligence and workflow automation.

This role is hybrid.

Responsibilities

• Architect and develop a local, GPU-optimized AI platform for document intelligence and RAG/agentic systems.
• Develop scalable pipelines to extract structured data (tables, diagrams, text) from vendor documents using OCR and vision-based parsing.
• Implement RAG pipelines, vector search, and agent-based AI workflows, ensuring outputs are explainable and verifiable.
• Build and maintain a user-facing, on-prem AI application for document upload, data review, and workflow triggering.
• Optimize model inference and system throughput on the onsite GPU cluster by managing batching, memory use, and I/O.

Required Skills

• 12+ years of overall experience, including 5+ years in AI/ML, NLP, or Data Engineering.
• Strong experience with LLMs, including open-source models, fine-tuning, and quantization.
• Proficiency in Python, PyTorch, and TensorFlow.
• Demonstrable experience building end-to-end pipelines (ingest → extract → transform → deliver).
• Hands-on experience deploying AI models on on-prem GPU infrastructure.
• Strong experience with RAG systems and vector databases (e.g., FAISS/Milvus/Chroma).
• Familiarity with agent frameworks (LangChain, LlamaIndex).
• Experience with OCR, document extraction, or layout models.

Preferred Skills

• Experience with layout models for complex document structures.
• Background in data engineering for large-scale document processing.

Key Skills

Education

Any Gradute