Skip to main content
Sandip M.c.

Sandip M.c.

Research Assistant

Talent
Hattiesburg, MSMember since November 18, 2026
PythonSQLPyTorchHuggingFace TransformerspandasReactFlaskREST APIsD3.jsPostgreSQLSQLAlchemy
·

Hire this person

Send a job offer directly to this candidate

PythonSQLPyTorchHuggingFace Transformerspandas

About

Conducted systematic literature review analyzing 40+ research papers on LLM fine-tuning techniques and prompt engineering, synthesizing findings to guide experimental design for domain-specific model adaptation in healthcare and bioinformatics Implemented NanoChat LLM architecture from scratch using PyTorch, conducting 15+ training experiments to quantify impact of data ordering, historical sequencing, and text perspective on model convergence speed and output quality Designed reproducible experimental framework with automated evaluation metrics to assess model performance across varied training configurations, reducing manual testing time by 60% and establishing foundation for scalable fine-tuning pipeline

Experience

Research Assistant - Department of DATA SCIENCE, Mississippi State University - Starkville, MS

(2025-09)

  • Conducted systematic literature review analyzing 40+ research papers on LLM fine-tuning techniques and prompt engineering, synthesizing findings to guide experimental design for domain-specific model adaptation in healthcare and bioinformatics
  • Implemented NanoChat LLM architecture from scratch using PyTorch, conducting 15+ training experiments to quantify impact of data ordering, historical sequencing, and text perspective on model convergence speed and output quality
  • Designed reproducible experimental framework with automated evaluation metrics to assess model performance across varied training configurations, reducing manual testing time by 60% and establishing foundation for scalable fine-tuning pipeline
  • Collaborating with faculty advisor to develop accessible fine-tuning interface that will enable non-technical researchers to customize LLMs for specialized applications without deep ML expertise

Research Intern - NAAMI (Nepal Applied Mathematics and AI Research Institution) - Kathmandu, Nepal

(2023-06 - 2023-10)

  • Developed ensemble machine learning models achieving 89% accuracy for disease prediction across 5,000+ patient records by integrating structured clinical data (demographics, lab results) with unstructured medical imaging (CT, MRI scans)
  • Engineered deep learning segmentation pipeline using U-Net architecture that reduced false negative rate by 23% in brain tumor detection from MRI scans, enabling earlier diagnosis for 150+ neuro-oncology cases during pilot deployment
  • Built automated data preprocessing pipeline handling 10GB+ of multimodal cancer omics datasets, reducing data preparation time from 8 hours to 45 minutes and standardizing quality control across 12 research projects
  • Created interactive visualization dashboards using Python and Plotly that enabled clinical research teams to explore patient stratification patterns, resulting in identification of 3 novel biomarker candidates

Education

Bachelor of Science - Artificial Intelligence - Mississippi State University (2025-08 - 2029-05)

Skills

Research Assistant

Reviews

Related