William Steimel

•Led the technical development and oversight of Commercial Pharma AI systems in Japan and globally, directing a team of 8 engineers and data scientists in executing their work.

•Architected and launched a next-gen ML and data engineering platform on Databricks, optimizing automation and ML model deployment. Reduced pipeline runtimes by 80%, enabling daily automation and saving ~$100K annually in compute and labor costs.

•Engineered an agentic generative AI web app using React, Next.js, TypeScript, Python, FastAPI, and AI content generation APIs, leading end-to-end development and enhancing global content generation capabilities.

•Deployed and scaled an MVP Omnichannel Decisioning AI engine, expanding usage from 0 to ~2000 users across multiple brands and markets. Delivered multiple ML system deployments, driving significant impact across priority brands while reducing technical resource costs and saving £1.5M in outsourcing.

•Served as the key point of contact for data requests, interfacing with business stakeholders in Japanese, prioritizing tasks, and facilitating agile processes as Scrum Master to ensure timely and high-quality delivery.

•Reduced vendor lock-in by upskilling 4 internal data scientists and engineers for smoother, higher-quality delivery.

•Fostered a knowledge-sharing culture by maintaining comprehensive documentation and best practices, equipping new members with essential Japan Pharma and technical expertise.

•Led the development of multiple successful PoCs, generating interest in genAI and securing significant funding.

•Reduced 15-25 hours of manual development work per document layout to one UI click, through development of scalable Azure Machine Learning Pipelines for self-service training of layout based NER models for extraction of key-value pairs from financial documents.

•Utilized Scrapy web scraping and crawling framework and developed reusable ingestion scripts on over 20 Japanese open datasets for KPMG’s global Signals Repository platform.

•Maintained code quality and documentation of pipelines and software libraries for easier onboarding of new members.

•Designed and developed configuration based excel extraction and spreadsheet analysis library with open source libraries Pandas/Openpyxl/xlrd for ingestion of real estate property management reports and other related documents.

•Designed, planned, and implemented six generic reusable python libraries utilized across KPMGs document ingestion platforms.

•Implemented generic feature extraction functions on polygon shape data utilizing vector and raster data. Functions contributed to geoscaler python library for scaling geospatial data between geographical levels.

•Developed scenario based simulation system with Streamlit and Docker to forecast profit loss and cash flow of buildings in REIT portfolio using accounting and real estate domain knowledge.

•Analyzed data from internal affiliate marketing application to get insight into partner behavior and develop recommender system.

•Responsible for management of overall quality of IT Service Areas including Data Center Site Mgmt, IT Hosting Ops, Critical Incident Response Team, Enterprise Change Control, and Mainframe

•Responsible for implementing reporting automation and statistical analysis to facilitate process optimization, capacity/demand forecasting, service level management, and audit compliance