Lance J.

About Lance J.

I don\'t just train models; I architect the governance that makes them…

Experience

1. Ai Resident

   Apziva

   May 2025 — Present · Istanbul, TR

   Architected & Delivered End-to-End ML Systems: Owned the complete project lifecycle for mission-critical platforms, including a real-time fraud detection system that reduced project delivery time by an estimated 40%. Architected and built a unified streaming analytics platform using a resilient Kafka/Spark pipeline that processed 100K+ events/min and achieved a 0.90 F1-Score in identifying fraudulent transactions in under a second.Developed Impactful Predictive Models: Engineered a machine learning model to predict customer happiness, achieving 73% accuracy and providing actionable insights that led to operational improvements. Implemented a predictive model for a fraud prevention platform that reduced false positives by 25%, directly improving customer experience.Engineered NLP-Powered Talent Solutions: Created an NLP-powered talent discovery engine that leveraged vector embeddings and a human-in-the-loop relevance feedback mechanism to semantically rank candidates. This system achieved a 90% reduction in initial screening time by automating the identification of top-tier profiles from a pool of 100+ candidates.Designed & Optimized Data Pipelines: Developed robust data pipelines using Python and Pandas to ingest, clean, and process data from diverse sources, ensuring 100% data availability for downstream machine learning models.Cross-Functional Collaboration: Collaborated with cross-functional teams to address operational challenges and align data solutions with global expansion strategies in a fast-growing startup environment.