NLP Specialist — Legal Named Entity Recognition

We build contract analytics software. Our current NER pipeline was fine-tuned 2 years ago on a limited dataset and it's not keeping up with contract complexity and variety we now see. Specifically underperforming: party name extraction across non-English names, obligation vs. permission vs. prohibition classification, date expressions in non-ISO formats, and nested entity references. We need an NLP specialist to audit what's failing, build a better training dataset with our annotation team, retrain/fine-tune (likely SpaCy + transformer), and benchmark against our test set. Output: improved NER model, evaluation report, annotation guidelines update.