Chicago, Illinois, United States
TensorFlow, PyTorch, TFX, Ray, Kubernetes, XGBoost, LLM-as-a-Judge, GenAI, MLflow
• Improve the sales outcome by developing a recommendation graph algorithm used by 2,100 sales agents using Python, AWS Sagemaker, Flask, Docker, and Jenkins, increasing conversion rate by 7% and retention rate by 5% • Optimize sales-call-queue prioritization system by improving ROCAUC of scikit-learn ensemble model by 2% and conducting MLflow-based A/B tests in production, driving 7% lift in conversion rate and $4M in revenue • Implement a customer care priority ranking algorithm using a LightGBM classifier that predicts the likelihood of customer attrition, generating over $5M in revenue through improving customer retention • Design end-to-end system architectures for Machine Learning products by leveraging object-oriented design, Python, AWS (S3, Sagemaker, Athena), PySpark, Docker, Flask, Gunicorn, Jenkins, and New Relic • Improve model latency of an app that handles 30,000 daily requests by optimizing the algorithmic complexity, lowering the response time by 40% • Enhance marketing effectiveness by designing customer segmentation using DBSCAN and creating marketing targets based on predicted ad responsiveness and customer look-alikes using LightGBM classifiers • Plan Machine Learning (ML) Roadmap by collaborating with business partners to identify ML product opportunities and ensuring alignment of ML projects with business goals, achieving 100% product adoption • Manage 2 Data Scientists and facilitate professional growth by directing regular one-on-one meetings, code-reviews, and learning sessions • Lead Data Science recruiting efforts by coordinating with the recruiting team to identify prime candidates and conducting over 100 interviews • Skills Used: Python, Scikit-Learn, LightGBM, AWS, Flask, Docker, Git, Pandas, NumPy, PySpark, Seaborn, Splunk, New Relic, Jenkins, Bash
• Developed a browser-based app that examines and visualizes the relationship between economic indices and campaign KPIs using Python, Plotly, and APIs, automating 10 hours of weekly manual analysis efforts • Analyzed hypothesis tests and recommended data-driven advertising strategies to clients, increasing conversion rate by 10% • Determined optimal ad-serving frequency by analyzing event-level data in Hadoop using R and Hive, improving advertising efficiency by 25% • Identified the contribution of different advertising channels by creating logistic regression models using PySpark and Shell, informing budget allocation decisions and Multi-Touch Attribution analysis • Skills Used: Python, R, SQL, Scikit-Learn, Pandas, Plotly, AWS, Hadoop, PySpark, Tableau, Shell, Hive
• Analyzed historical data and hypothesis tests to recommend data-driven advertising strategies to clients • Automated the weekly reporting process using PostgreSQL and Excel, improving the accuracy and efficiency of weekly reporting by 50% • Collaborated with agency and internal teams to ensure database quality for 60+ digital ad campaigns • Skills Used: SQL, Excel, PowerPoint