SFT trainer - for client NIVIDA, Turning
I generated and refined supervised fine-tuning (SFT) data for large language models for Nvidia via Turning. I evaluated model responses for adherence to instructions, reasoning quality, and accuracy. I played a key role in preparing datasets for training advanced language models. • Created prompt and response datasets for model training • Rated LLM outputs based on provided benchmarks • Offered feedback on model accuracy and coherence • Supported the SFT process with detailed evaluations