Abby N. Mogusu

About Abby N. Mogusu

I build clinical AI systems that turn complex healthcare data into actionable decisions: reducing delays, improving outcomes, and enabling smarter care delivery at scale.As a Machine Learning Engineer with 2+ years of experience across healthcare and production ML systems, I specialize in designing and deploying end-to-end AI solutions that operate in real clinical and enterprise environments. My work sits at the intersection of clinical operations, machine learning, and scalable infrastructure. In healthcare settings, I’ve developed models on clinical procedures that reduced prediction error by 50% and improved scheduling accuracy, helping care teams minimize delays and optimize resource allocation. By combining structured data with clinical NLP, I’ve improved model performance (AUC 0.78 to 0.86) while reducing manual chart review effort at scale. Beyond modeling, I focus on deploying production-ready systems. I’ve built and scaled ML pipelines using AWS (SageMaker, Bedrock, Lambda) and modern MLOps practices enabling faster experimentation cycles (3x) and reliable model deployment in high-impact environments.What I Specialize In1. Clinical NLP, predictive modeling, decision support systems, EHR data analysis, healthcare operations optimization2. Generative AI for Healthcare- RAG pipelines, LLM applications, semantic search, medical & regulatory document intelligence3. Production ML & MLOps- AWS (SageMaker, Bedrock, Lambda), CI/CD, model deployment, monitoring, scalable inference4. End-to-End ML Engineering- Experimentation - model development - deployment - production optimizationLet’s Connect If You’re- Building clinical AI / digital health / health-tech products - Scaling ML systems in hospital, biotech, or pharma environments - Exploring LLMs, NLP, or decision support systems in healthcare

Experience

1. Ai Engineer, Pfizer

   Extern

   Feb 2026 — Present

   Pharmaceutical companies like pfizer manage thousands of compliance documents. A single outdated file, buried in a 100+ page vendor bundle, can trigger FDA warning letters, production delays, and recalls. The industry is racing to automate document processing and how AI can catch what humans miss.• Contributed to an AI-powered document intelligence platform for pharmaceutical vendor compliance by building OCR and RAG pipelines with Tesseract, PaddleOCR, LlamaIndex, Mistral, and Phi-2, enabling automated extraction and semantic search across regulatory documents.• Designed a retrieval pipeline that combines vector search with LLM reasoning to cut manual document review effort and improve detection of outdated or FDA-risk compliance documents across vendor submissions.• Built a user-facing compliance review interface that lets stakeholders query vendor documents in natural language, speeding up regulatory analysis and streamlining internal audit workflows.

Education

• Michigan State University

  Master's degree, Data Science