C
Chris Harry Patrick
About
Detail
Milwaukee, Wisconsin, United States
Résumé
Jobs
verified_user
0% verified
AI/ML EngineerUnitedHealth GroupJan 2025 - Current (1 year 8 months)• Improved patient risk scoring accuracy 28% using AWS SageMaker models, reducing readmissions 15% and supporting healthcare decision-making nationwide. • Reduced data latency 40% by automating ETL workflows with Python, Pandas, SQL, and AWS S3 across healthcare systems. • Ensured HIPAA compliance by integrating fairness monitoring frameworks, strengthening Responsible AI adoption across enterprise level machine learning healthcare applications. • Boosted workflow efficiency 25% by deploying scalable Flask and FastAPI inference APIs, enabling real-time healthcare decision support. • Delivered actionable insights by developing reporting dashboards integrating predictive model outputs, improving executive healthcare decision-making processes- AI/ML EngineerHexaware TechnologiesJul 2022 - Aug 2023 (1 year 2 months)• Reduced credit risk prediction errors 22% by building regression and classification models, enhancing financial portfolio quality significantly. • Decreased manual review workload 35% by implementing CNN and RNN architectures using TensorFlow and PyTorch for NLP. • Achieved 99.5% uptime by deploying Docker-based MLOps pipelines with CI/CD workflows across enterprise AI applications. • Accelerated decision-making by embedding machine learning outputs into Power BI dashboards, providing executives with actionable business insights. • Increased scalability through automated retraining workflows, improving model lifecycle management and reducing deployment overhead for clients. • Accelerated reporting speed by applying deep learning architect
- ML EngineerHexaware TechnologiesJun 2021 - Jun 2022 (1 year 1 month)• Increased e-commerce engagement 15% by developing recommendation systems that improved personalization and optimized targeted marketing campaigns. • Maintained enterprise model accuracy by validating clustering, decision tree, and classification techniques for robust deployment. • Enabled informed decisions by creating Tableau and Power BI dashboards presenting machine learning-driven insights. • Reduced experimentation cycles 30% by optimizing workflows using Jupyter Notebook and Google Colab environments effectively. • Improved personalization by applying clustering algorithms to customer behavioral data, generating actionable insights. • Enhanced marketing campaign performance by integrating predictive machine learning outputs into str
Education
verified_user
0% verified
- UMaster of Science in Computer ScienceUniversity of Wisconsin, MilwaukeeMay 2025
- VBachelor of Engineering in Computer Science and EngineeringVelammal Engineering College, ChennaiApr 2023Affiliated to Anna University, Chennai, India
- Scrum: The BasicsLinkedIn Learning
- AWS Certified Cloud Practitioner (CLF-C02)LinkedIn Learning
Projects (professional or personal)
verified_user
0% verified
- AI PDF Chat ReaderIndependent• Built a privacy-preserving PDF Q&A app end-to-end: Next.js/React UI with SSE token streaming, Express/TypeScript API, LangChain RAG over HNSWLib, nomic-embed-text embeddings, and local Mistral 7B via Ollama; ingestion pipeline parses PDFs, splits ~1k-char chunks (200 overlap), and returns grounded answers with page-level citations. • Improved retrieval quality and reliability with whole-document “summarize” mode, deduped citations, EventSource→fetch fallback + heartbeats for robust streaming, and an accessible Tailwind UI; supports multi-PDF uploads and runs fully offline at zero API cost.
- Federated Learning for Image ClassificationIndependent• Boosted model accuracy 12% and reduced convergence time and communication costs 18% by engineering a federated learning pipeline with adaptive client participation and smart aggregation strategies on Azure. • Simulated non-IID data distributions to replicate real-world heterogeneity, strengthening robustness of distributed training and ensuring scalable, privacy-preserving machine learning deployment across heterogeneous client environments.
- Personal Portfolio WebsiteIndependent• Built a responsive portfolio site with React and Tailwind CSS, emphasizing accessibility, performance, and clean UI to showcase AI/ML projects, publications, and open-source contributions. • Increased recruiter engagement by centralizing technical achievements, publications, and demos into a professional online portfolio, supporting visibility and communication of applied engineering expertise.
Publications
verified_user
0% verified
- FFederated Learning for Image Classification with Dynamic Data and Adaptive ClientsJan 2025