I'm Daniel Palacios, a 4th-year Ph.D. Candidate in Quantitative & Computational Biosciences at Baylor College of Medicine, and an NSF Graduate Research Fellow (2024) and NLM Fellow (2023). My research focuses on building LLM-based AI agents, AutoML systems, and predictive models for clinical informatics and healthcare applications.

• Developed hybrid pipelines combining AutoML with LLM-based agents to predict postpartum depression using clinical notes from first hospital visits.

• Used AI agents and statistical modeling to analyze links between sleep disorders and epilepsy in pediatric patients.

• Built LLM-powered agent pipelines with retrieval-augmented generation (RAG) to extract HPO terms and identify rare disease cohorts from unstructured clinical data.

• Engineered LLM-based tools for phenotype extraction from clinical notes, validated against clinician annotations.

• Applied AutoML and agent-based inference systems to identify DKA risk factors in children with Type 1 Diabetes.

• Created a semantic similarity AI agent using RAG to match patients in large clinical databases, enhancing decision support.

• Languages/Frameworks: Python, PyTorch, scikit-learn, HuggingFace, Streamlit
• LLM Techniques: QLoRA fine-tuning, RAG, PEFT
• Cloud: AWS SageMaker, Amazon Bedrock
• Tools: Git, Docker, Conda, Jupyter, VS Code

Teaching Assistant – QCMB 1 & 2 (2023–2025)

• Supported instruction in Quantitative & Computational Methods for Biosciences
• Led lectures, graded assignments/exams, and created supplemental materials on Probability Theory, Statistics, and ML for Biosciences

• 📫 Email: daniel.palacios@bcm.edu
• 💼 LinkedIn: https://www.linkedin.com/in/daniel-palacios-506734224/

Thanks for visiting! 🚀 I'm always open to collaboration and mentorship opportunities in AI + healthcare.