Sultan Abdul Qadir Uneeb Email & Phone Number

San Francisco Bay Area

• Identified new data sources for model improvement, and created an ETL pipeline in Apache Airflow. These pipelines downloaded, cleaned, transformed, and loaded 20+ million data points into an AWS RDS database every month.
• Used new data sources to create XG Boost models that provided over a 10% improvement relative to baseline claim prediction models.
• Developed automated report creation for customer POCs. The reports analyzed over 1,000,000 data points, drew relevant charts, calculated summary statistics, and provided key metrics for customers to use our product data.
• Developed Quality Control processes to QC and verify schemas for 40+ GB Geospatial data in under a second. This reduced QC time by over 40%.
• Maintained a $4 million revenue Casualty Insurance API with Python, Jenkins, SQL and a Postgres database. Improved API, performance by 20% via bug fixes in Python, Docker images, and AWS Sagemaker endpoints.
• Developed an automated data testing framework using AWS Sagemaker. The framework would process new data points then automatically make machine learning models to see if the data points improved performance.

San Francisco, California, United States

• Worked under Professor Diana Goldemberg (Economist at World Bank) to revise and develop course components for future economics students.
• Filtered income inequality data on over 120 countries, and coded scripts, using Python and Jupyter notebooks, that carried out automated regressions, summary statistic calculations, and gini coefficient calculations.

Islāmābād, Pakistan

• Used the NetworkX library and Python to develop an Agent-Based Network Model to predict the spread of COVID-19 in Pakistan’s districts.
• Provided information to officials in the army so they could order the relevant amount of PPE.
• The model also predicted which districts would get hit the hardest, and the predictions were ~80% accurate.

San Francisco, California, United States

• Assessed 2000+ undergrad applications for different programs over the course of three years while analyzing over 6000 data points to correctly quantify candidate abilities.
• Parsed Technical Documents in multiple languages such as transcripts and school profiles. Evaluated Applicants’ spoken and written English while gauging candidate data such as tests, interviews, essays, and life.

Yeongdeungpo-gu, Seoul, Korea

• Analyzed over 100,000+ rows of data to gain insights and devise new solutions.
• Designed a recommendation model for 500,000+ users based on a KNN Algorithm and a classification tree. Both models were developed purely using Python and the scikit-learn library.
• Presented the product to stakeholders interested in the project at Civitas (Minerva Schools at KGI event for future project partners). The presentation motivated other companies to offer projects and internships to.