Imagine a world where complex medical documents transform into concise, personalized summaries for everyone involved—doctors, patients, and the public alike. That's the exciting promise of persona-based summarization explored in this research. Traditional approaches to summarizing medical information often struggle with the high cognitive load and variability among human summarizers. Generic Large Language Models (LLMs), while powerful, can be costly and may lack accuracy without specialized training. This research introduces a two-fold solution. First, it demonstrates how to efficiently fine-tune smaller, domain-specific LLMs using a readily-available health corpus. This method makes personalized summarization more efficient and accessible. Second, the researchers employed AI-based critiquing to ensure the quality of the generated summaries. This approach was validated by comparing the AI's judgment with human assessments. The results are promising, showing high agreement between the AI and human critiques, especially with the fine-tuned Llama2-13b model. This model stood out for its ability to generate accurate, persona-specific summaries, while also being cost-effective. The implications of this research extend beyond healthcare. This efficient and scalable AI pipeline can be adapted for legal, corporate, educational, and other specialized fields. As the volume of domain-specific information continues to grow, the ability to create personalized, easily digestible summaries will be indispensable.