Senior AI/ML Engineer at Edvantis

Designed and deployed scalable AI platforms for healthcare and fintech, improving predictive accuracy and reducing inference latency by optimizing PyTorch and ONNX Runtime workflows. Fine-tuned LLMs (GPT-4, LLaMA 2, Mistral) for summarization, domain-specific retrieval, and enterprise conversational agents, ensuring consistent performance across multilingual use cases. Built automated ETL and feature-extraction pipelines using Airflow, Pandas, Polars, and OCR, enabling stable ingestion of unstructured documents at large scale. Created drift-monitoring and model-refresh systems using MLflow, DVC, and Evidently AI, allowing early detection of performance decay in production. Implemented privacy-preserving data workflows with synthetic data generation and automatic PII masking to maintain GDPR and HIPAA compliance. Developed scalable inference services using FastAPI, TorchServe, and Kubernetes, supporting low-latency model deployments across multi-cloud environments. Collaborated with Dev

Senior Data Scientist at Take2

Developed predictive models for fraud detection, churn scoring, and credit risk assessment using LightGBM and scikit-learn, improving accuracy and stability in production. Created SHAP-based explainability dashboards in Streamlit that allowed compliance and business teams to understand model decisions during audits. Integrated ML models into real-time pipelines using REST APIs and Kafka, enabling automated decisioning and reducing manual review time. Built Airflow-driven retraining workflows that refreshed models automatically, ensuring data drift did not degrade performance. Engineered feature-generation pipelines using Pandas and SQL that improved model input quality and reduced preprocessing time across large datasets. Collaborated closely with product and risk teams to translate business requirements into measurable data-science solutions with clear KPIs and deployment criteria. Deployed production models with MLflow and Docker, ensuring reproducible builds, controlled versioning,

Machine Learning Developer at Ascendis

Built analytics-driven applications and internal tools that powered segmentation, forecasting, and customer-behavior insights for marketing and operations teams. Developed data-processing components and ML-ready datasets using Pandas and async Python, improving speed and reliability of downstream modeling tasks. Implemented intelligent API middleware with retry logic, caching, and fallback handling, ensuring stable delivery of data to analytics and ML pipelines. Designed modular data-exploration interfaces with virtualized tables, server-side filtering, and optimized rendering for large datasets. Created automated test suites including E2E, smoke, and contract tests to guarantee smooth deployments and reduce regressions. Collaborated with UX designers to build responsive dashboards optimized for mobile and tablet, improving field-team adoption.