Michael Stringer

About Michael Stringer

Extremely curious and meticulous when given a challenging task. Believe that nothing is impossible, it only requires more research to overcome obstacles. Possessing an understanding of many complex subjects from advanced thermodynamics to deep reinforced learning. Combined with excellent team leading skills and management experience makes an excellent candidate for the task at hand.

Experience

1. Director, Machine Learning Developer and Data Scientist

   GSK

   Jun 2025 — Present · Colorado Springs, CO, US

   Architected AI-Programmer, a six-phase clinical data processing pipeline that orchestrates multi-LLM cascade (Gemini → Claude) to automatically generate production R code from ADaM/SDTM specifications—replacing weeks of manual programmer effort per study• Designed domain-aware validation system that auto-detects ADaM structure (ADSL/BDS/OCCDS) and generates appropriate comparison keys, achieving ≥90% match rate threshold required for regulatory compliance• Implemented three-tier caching architecture (file-level, LLM response, dataset) reducing pipeline costs and eliminating redundant API calls across 100+ document summarizations per study• Built self-improving feedback loop that analyzes failures against ground-truth SAS code, differentiates prompt vs. specification issues, and automatically synthesizes improved prompts for subsequent runs• Developed \"prompt-cascade\" architecture where fast/cheap models (Gemini Flash) perform binary relevance filtering before expensive models (Claude) generate code—optimizing cost per variable while maintaining accuracy• Created intelligent pipeline that converts clinical PDFs/DOCX/Excel to markdown, summarizes with domain-specific prompts, analyzes variable dependencies, and generates validated R code with admiral/metacore integration• Collaborate with AI engineers on model deployment and with cross-functional teams in preclinical research, clinical operations, and regulatory science to ensure scientifically sound solutions

Education

• The University of Texas at Arlington

  Bachelor of Science - BS, Mechanical Engineering

  2011 — 2014

• Udacity

  Deep Reinforcement Learning, Artificial Intelligence

  2019 — 2019

• Udacity

  Deep Learning, Artificial Intelligence

  2018 — 2019