Ken Oben Yoshimoto

Ken Oben Yoshimoto AI Solutions Architect & ML Engineering Leader SUMMARY Enterprise AI architect delivering 40-65% operational cost reduction through production-grade LLM and ML systems across regulated industries. Demonstrated success in transforming business operations with AI solutions that generated $10M+ ROI for Fortune 500 clients while maintaining 100% regulatory compliance. Skilled in bridging technical innovation with business strategy, leading crossfunctional teams through complex AI implementations, and communicating technical value to C-suite stakeholders. TECHNICAL EXPERTISE AI & ML ▪ ▪ ▪ Large Language Models: Production LLM application architecture, fine-tuning, RAG systems (94% retrieval accuracy) Machine Learning: Enterprise-scale ML pipelines, computer vision, NLP, predictive analytics Frameworks & Tools: PyTorch, TensorFlow, Hugging Face, LangChain, scikit-learn, vector databases (Pinecone, Weaviate) Cloud Architecture & Infrastructure ▪ ▪ ▪ AWS: SageMaker, EKS, Lambda, ECR, S3, CloudFormation, Step Functions, CloudWatch, IAM Other Platforms: GCP (Vertex AI), Azure (ML Studio), Databricks Containerization: Docker, Kubernetes, Helm, microservices architecture, serverless deployment Software Development ▪ ▪ ▪ Languages & Development: Python, JavaScript/TypeScript, SQL, Java, CI/CD, test automation DevOps: CI/CD (GitHub Actions, Jenkins), GitFlow, infrastructure as code, test automation (pytest, Jest) Database & Storage: PostgreSQL, MongoDB, Redis, Elasticsearch, vector databases, data optimization Leadership & Strategy ▪ ▪ ▪ Team Leadership: Technical mentorship, agile methodologies (Certified ScrumMaster), resource optimization Business Impact: Cost reduction strategies, ROI measurement, executive stakeholder engagement Regulatory Expertise: Financial compliance (SEC, FINRA, BSA/AML), Healthcare (HIPAA), data privacy (GDPR) PROFESSIONAL EXPERIENCE Principal AI Solutions Architect | Capgemini| NY, US (Remote) Apr 2020 - Present Financial Service AI Transformation (JP Morgan Chase) • • • • • Delivered $2.3M annual savings by engineering enterprise RAG platform that improved document retrieval accuracy from 71% to 94% while accelerating processing speed by 67% Reduced development cycles by 30% while leading 5-person agile team developing specialized chunking algorithms and hybrid retrieval systems that maintained 100% SEC/FINRA compliance Automated 65% of manual loan processing workflows through production-grade LLM applications using LangChain, OpenAI APIs, and custom prompt templates Protected millions in transaction volume with fraud detection system achieving 96% accuracy and 28% reduction in false positives using ensemble ML methods and anomaly detection algorithms Generated $7.8M in validated annual savings by translating technical AI initiatives into business outcomes in bi-weekly Csuite presentations Healthcare Analytics Platform (Pfizer) • • • Reduced patient readmissions by 18% through HIPAA-compliant ML pipeline built with TensorFlow, PyTorch, and Databricks, significantly improving clinical outcomes across 12 facilities Achieved SOC 2 Type II audit compliance by implementing comprehensive security frameworks enabling secure crossfunctional data collaboration Delivered 99.97% uptime and 42% cost reduction by optimizing cloud infrastructure through strategic resource allocation and performance monitoring Senior Software Engineer | Santa Claus House | North Pole, Alaska Oct 2017 - Mar 2020 • • • • Generated $450K additional annual revenue by leading development of hybrid recommendation engine using TensorFlow, AWS SageMaker, and collaborative filtering that increased average order value by 32% Reduced support costs by $120K annually by pioneering AI-powered customer service platform with 95+ intent recognition capabilities, resulting in 28% improvement in customer satisfaction Maintained sub-second response times during 20x traffic spikes by implementing serverless architecture with AWS Lambda, DynamoDB, and auto-scaling mechanisms Decreased inventory stockouts by 30% through predictive inventory system with time-series forecasting models, improving ordering accuracy by 42% and optimizing $3.2M in annual inventory Software Engineer | Banner Health | Tucson, AZ, US • • • May 2014 - Sep 2017 Achieved 22% reduction in 30-day readmissions by building predictive patient risk system using HIPAA-compliant ML pipelines, saving approximately $350K in penalty avoidance Detected anomalies with 88% sensitivity by developing medical imaging analysis platforms using convolutional neural networks and transfer learning, augmenting radiologist capabilities Improved physician compliance by 34% through clinical decision support system that reduced inappropriate test ordering by 22%, enhancing quality metrics across 8 hospitals SIGNATURE PROJECTS Financial Compliance LLM Evaluation Framework • • • Developed industry-leading benchmark achieving 92% accuracy in measuring regulatory compliance for financial text analysis Open-sourced framework (450+ GitHub stars, 120+ forks) adopted by three major financial institutions, establishing new industry standard for responsible AI validation Published technical whitepaper on methodology receiving recognition from financial regulatory bodies Enterprise Banking Document Processing Pipeline • • • Architected end-to-end system with custom NER models reducing manual processing time by 86% and error rates by 74% Integrated with legacy banking systems through secure APIs meeting SOC 2 and FINRA audit requirements Scaled to process 35K+ documents daily with 99.9% availability and full auditability Risk Assessment Analytics Dashboard • • • Created interactive risk visualization platform combining multiple data streams with XAI techniques including SHAP and LIME Deployed across 200+ risk analysts with regulatory-compliant decision rationales, increasing risk identification by 42% Reduced time-to-decision by 67% while improving consistency of risk evaluations by 53% EDUCATION BACHELOR OF SCIENCE | Computer Science | University of Alaska Fairbanks Aug 2010 - May 2014