Humza Haider Email & Phone Number

Minneapolis, Minnesota, United States

Minneapolis, Mn

Founded a Software-as-a-Service personalization platform powered by contextual multi-armed bandits. • Designed and implemented the platform from scratch utilizing Golang, Terraform, and GCP technologies, including Google Kubernetes Engine (GKE), BigQuery, Pub/Sub, Cloud SQL.• Actively participated in customer meetings, showcasing our platform’s capabilities and addressing clientquestions and concerns.

Minnesota, United States

• Spearheaded the design and implementation of an end-to-end platform for Contextual Multi-Armed Bandits, optimizing decision selection and ranking. This platform saw the following achievements:(1) Attained a 40% increase in sponsored product ad revenue ($2.6M annualized) by dynamically selecting optimal arrangements of featured products in real time.(2) Orchestrated a $1.7M annual growth in order revenue by personalizing category rankings on the homepage.(3) Enhanced customer engagement with new retailers by 22% through personalized retailer recommendations.• Marketing Decision Engine: Designed and implemented a comprehensive platform mediating between digital marketing and internal promotions across the website and app. Balanced manual constraints (budgets, impression rules, seasonal events) while tailoring content based on user shopping profiles, purchase history, and site context using machine learning.• Tech Lead: Led a team of 5 developers on implementing the multi-armed bandit platform. Acted as the primary designer, developer, and reviewer of all components on the platform. Mentored and guided junior developers to improve their technical skills and holistic understanding of the platform.

Eden Prairie, Minnesota, United States

• Using Kafka, streamed thousands of predictions per second identifying late shipments before they occurred. This led to better customer outcomes by informing the customers preemptively to compensate for late arrivals and alerted warehouse teams to accept shipments later than previously anticipated.• Created Kafka producers to stream model predictions and ground truth events into snowflake for data science and analytical workloads. Monitoring was built for real time model performance as well as overtime analyses to review model outcomes.

Greater Minneapolis-St. Paul Area

• Prepared, trained, and deployed a multi-task deep learning model using 200M+ training instances that identified potentially late truckload shipments up to days ahead of time, improving over previous models by a factor of 3 in MAE. This model empowered internal representatives to then work with customers to adjust their subsequent supply chains.• Communicated directly with business stakeholders to help design custom solutions while maintaining scalability. Such solutions helped to uphold the relationship with multi-million dollar customers and simultaneously save the business thousands of dollars by keeping within service level agreements.

Edmonton, Alberta

• Evaluated Multi-Task Logistic Regression, Random Survival Forests, Cox Proportional-Hazards, and Accelerated Failure Time survival models across a variety of evaluation metrics for time to event data.• Constructed novel methodology to model and evaluate patient specific survival prediction.• Built an R package to support Multi-Task Logistic Regression.• Estimated customer reservation prices using survival prediction techniques andcommunicated how these estimates can lead to increases in profit.• Generated manuscript for peer-reviewed publication (under review at Journal of Machine Learning Research). • Advised and supported a student group project whose work was accepted to IJCAI 2019 (Simultaneous Prediction Intervals for Patient-Specific Survival Curves).

Alberta, Canada

• Led practice seminars to give students supplemental class material and practice problems.• Held office hours to give student support for assignments/quizzes/exams and general questions.• Graded assignments and gave additional feedback to help students overcome academic challenges.

San Antonio, Texas

• Developed statistical models to (1) identify important features for prediction of dental emergencies, (2) conduct a meta-analysis of dental emergency rates in deployed Sailors and Marines, and (3) examine the usage of new dental technology among Navy and Marine Corps dental treatment facilities. • Consulted with teams and individuals on their research projects within the dentistry research unitand recommended appropriate statistical tests and suitable modelling techniques on a project specific basis.• Assisted in producing white papers, technical reports, articles, and presentations of technical material.

University Of Minnesota Morris

• Operated within a research team to analyze the effects of Obstructive Sleep Apnea (OSA) on commercial truck driver’s crash risk, crash cost, and medical care.• Data from medical forms and health insurance claims were aggregated to scrutinize the ability of medical examiners to identify OSA and hypertension.• Utilized STATA for data processing and analysis of millions of health insurance claims, trucking operational data, and other protected health information (PHI) data.• Assisted in manuscript generation, editing, and revision process.

La Crosse, Wisconsin Area

• Modelled survivability of the Indiana Bat species under different conditions utilizing simulations, matrix models, branching processes, and other growth models. • Developed simulation code in R (https://github.com/haiderstats/PBR-Allee).• Drafted manuscript for peer-reviewed publication in Natural Resource Modelling.

Master Of Science - Ms, Computer Science

Bachelor Of Arts (Ba), Statistics, Computer Science, And Mathematics