ML/AI Engineer (OCR & Document Structure)

About the role

You will own the document accuracy project at the core of a high‑volume financial document processing pipeline. The role involves scaling investigations, shaping the ML and LLM strategy, and delivering production‑ready solutions that improve how the platform extracts and understands complex financial documents.

Key responsibilities

• Define direction and drive measurable improvements in document accuracy for a large‑scale financial pipeline.
• Investigate production data at scale, identify systematic error patterns, and translate insights into high‑impact system changes.
• Design, build, and deploy ML‑powered solutions such as OCR correction, entity matching, extraction, and LLM‑assisted analysis.
• Run structured experiments, validate models, and collaborate with backend teams to integrate successful approaches into production.
• Maintain long‑term ownership of deployed models, monitoring performance on live traffic and iterating to sustain accuracy gains.

Required profile

• 6+ years of experience in machine learning or NLP on production systems.
• 2–3+ years focused on document‑related problems (OCR, information extraction, structured/unstructured processing).
• Strong forensic data investigation skills on large, messy datasets.
• Proven ability to implement robust measurement and validation practices.
• Experience delivering ML solutions to production, including integration of LLMs with clear cost/accuracy/latency trade‑offs.

Required skills

• Python
• SQL
• Machine Learning (ML) and Natural Language Processing (NLP)
• OCR and information extraction techniques
• Large Language Models (LLMs) integration
• Modern MLOps practices
• Cloud ML platforms (e.g., Azure ML)