LLM Fine-tuning Engineer
TextFine Tuning
Fine-tuned open-source LLMs on domain-specific corpora using LoRA/QLoRA and Hugging Face PEFT. Managed the full pipeline from data tokenization, cleaning, and preparation to supervised fine-tuning and evaluation. Used the refined models in downstream Generative AI and Retrieval-Augmented Generation applications.• Prepared custom domain text datasets by cleaning, tokenizing, and structuring for effective model training
• Performed supervised and parameter-efficient fine-tuning using Hugging Face Transformers and PEFT
• Conducted quality checks and model evaluation using BLEU, ROUGE, and perplexity metrics
• Deployed trained models as inference endpoints after end-to-end validation and benchmarking
Fine-tuned open-source LLMs on domain-specific corpora using LoRA/QLoRA and Hugging Face PEFT. Managed the full pipeline from data tokenization, cleaning, and preparation to supervised fine-tuning and evaluation. Used the refined models in downstream Generative AI and Retrieval-Augmented Generation applications.• Prepared custom domain text datasets by cleaning, tokenizing, and structuring for effective model training • Performed supervised and parameter-efficient fine-tuning using Hugging Face Transformers and PEFT • Conducted quality checks and model evaluation using BLEU, ROUGE, and perplexity metrics • Deployed trained models as inference endpoints after end-to-end validation and benchmarking
2024 - Present