Dequan Er

Lead Data Scientist @National Debt Relief, LLC

NY, US

Spearheaded a lead recommendation system collaborating across sales, marketing, engineering teams to rank and dynamically route leads. This resulted in a proven 68% YoY growth in clients, a 92% YoY increase in enrolled debt, and a 55% surge in conversion rate post-deployment, significantly driving revenue and business expansion. • Engineered an end-to-end pipeline using Snowflake and MLOps tools, achieving 98% model accuracy in training, reducing inference latency to under 120 ms, and processing 1.5 M+ daily inferences with uptime. • Developed scalable RESTful APIs for real-time model inference within an MLOps framework using DataRobot, with automated retraining to maintain accuracy. • Led the development and deployment of time-series ARIMA), NLP analysis (LSTM for chat) and cancellation models, boosting ROI by 13% across channels and identifying $60MM in yearly revenue savings by reducing churn. • Led and mentored a team of junior data scientists, driving the development of machine learning strategies and fostering a culture of continuous learning. Collaborated closely with senior leadership to align firm-wide data initiatives with strategic business goals, ensuring impactful data-driven decision-making across departments.