Howard Small Email & Phone Number

Mountain View, California, United States

Tech Lead for the Memory Recommendation system. Spearheaded the use of ML-based optimization techniques to serve the best memories (i.e. auto-generated photo albums) to our users each day. - My team’s work contributed to a > 25% improvement in our north star metrics for user engagement. - Developed, deployed, and maintained ML models to identify a users most meaningful and relevance memories. - Pioneered the use of integer programming to select the optimal set of memories (subject to product constraints).

Mountain View, California

ML infrastructure engineer, implemented and managed services that computed > 10K ML features and predicted the fraud probability for 15K QPS of Google payment transactions. - Built a real-time data aggregation service which enabled serving 1.5M QPS of SQL query results in < 30ms. - Built a real-time graph mining service to identify related users and prevent repeat offenders using new accounts. - Built a historical replay pipeline for our entire stack to facilitate backfills of new features for ML model training (saving engineers 6 months of wait time).

This course introduces the fundamental concepts of Object Oriented Programming (OOP) through Java/Android-based coding projects. My responsibilities were:- Holding weekly office hours with students.- Grading and providing feedback on student project submissions.

Mountain View, Ca

Worked on the Payments Risk team which uses Machine Learning to identify and prevent potential fraud attacks for a number of Google products. I was assigned two projects: 1) Implement a “sanity check” tool to verify that ML models trained by our team had parameters that matched reasonable expectations based upon trends in the data.2) To streamline our ML pipeline by combining a number of disjointed and difficult to configure jobs into one simple script.

- Teach a weekly CS106 section, typically containing between 8 and 12 students (sectionees).- Grade sectionee assignments and hold Interactive Grading (one-on-one) sessions with each student to discuss grades, code/style, and how to improve.- Hold office hours in the LaIR for 2+ hours a week and help CS106 students with questions and debugging of code in both Java and C++.

Mountain View, California

"Google's Borg system is a cluster manager that runs hundreds of thousands of jobs, from many thousands of different applications, across a number of clusters each with up to tens of thousands of machines.” My intern project was do develop a machine learning (ML) system to predict when various Borg machines and/or jobs would behave abnormally. To do this I primarily leveraged Google’s Tensorflow API for Python to construct and train neural networks for various predictive tasks.

Mountain View, Ca

Worked on an internal event creation and management tool for Google hosted events. My primary project was designing and developing a survey application that allowed for the analysis of user sentiment regarding a particular event. This analysis is used to improve the overall quality of Google events. The entire application was written in Java, using Google’s Stubby RPC and Spanner database for the backend web service and Google Web Toolkit (GWT) for the frontend. Additionally, I contributed to design and implementation discussions for the overall event management system.

Herndon, Va

Worked for the Wearable Technology Initiative and helped to develop a platform which integrated the use of Smart/Wearable Android technologies to give users a seamless experience across multiple devices. My chief contribution was the development of an Android application using Java, RESTful Web Services, and a MongoDB database. The application allows managers to upload task-lists to the Android devices of direct reports and monitor their progress remotely. Also aided in the design of a MongoDB database model for the entire platform.

Arlington, Va

During my time with Lemur Retail, I developed an HTML Web Scraper (in C#) to gather information about all of the Nike products from the online Nike store and compile it into a CSV file. I was primarily supervised and guided by Shawn Hinsey, the Chief Architect of the Lemur Retail.

Master Of Science (M.S.), Computer Science, 4.05

Dual Track: Artificial Intelligence and Information Management/Analytics

Bachelor Of Science (B.S.), Computer Science, 3.82

Track: Artificial Intelligence

4.41

Activities and Societies: Black Student Union, FIRST Robotics Team 3455, Varsity Basketball, Fellowship of Christian Athletes, National.