Synthetic Training Data Generator for Domain Adaptation

ML engineers fine-tuning LLMs for specialized domains (legal, medical, financial) often lack sufficient labeled examples and cannot use proprietary data for training. This agent designs and implements synthetic data generation pipelines that produce diverse, high-quality instruction-following examples from a domain seed corpus.