Christian L.

I work on intelligent document processing systems using composite AI / machine learning.

• Began groundwork on automated recommendations for LIFE apps users • Worked with small team of 3 folks with architecture and implementation as well as technical updates to the company • Build back-end components with various AWS services in TypeScript and Python • Implemented and productionized self-training Decision Tree model for configurable page classification with 4-7x improved (0.46-0.96 F1) over existing implementation using MIMIC-III • Engineered and deployed self-training Deep Learning model (Keras; 25% better; 80% Acc; BCE 0.62; 0.15 Z-score Abs. Err.) for medical documents to segment and predict document date per page • Pioneered cost-effective Terraformed implementation of ML services at company with custom Python Docker image, AWS Lambda, and AWS EFS for model storage • Provided data and visualization support to inter-disciplinary team of scientists for intermittent fasting and retention analysis publication of +150K active mobile users • Developed Python SDK and assisted researchers in retrieving, analyzing, and exploring sizable real clinical and genomic data from the Precision Health Cloud • Redesigned/delivered intermittent fasting week/month/year views from data exploration to production for +2M users (l.rcd.zone/fasting-viz-release) • Organized regular AI/ML meetings for cross-team knowledge sharing • Developed full-stack features for security compliance and evidence collection in AWS Lambda / ECS • Performed privacy-respecting, automated ETL in TypeScript and Python for ML features • Crafted pipeline for automated ML model deployment with AWS Comprehend in Python • Built security GRC features in a Cloud-Native AWS environment for JupiterOne using TypeScript and Terraform • Created custom query language autocomplete system with statistically-based unit tests