Daniel Xu Email & Phone Number

San Francisco Bay Area

San Francisco Bay Area

Advanced Analytics & Machine Learning

San Francisco Bay Area

Built a framework to construct an optimization problem from retail customers' merchandise financial plans to determine optimal mix of choice counts while meeting revenue targets and reducing spending costs.- Iteratively experiment with customer data, identifying data issues and extracting reliable demand model and sell-through predictions- Supplied optimized plans to help guide customer merchandise buys to be more productive.

San Francisco Bay Area

Led design and implementation of multiple microservices backing Celect's scalable machine learning platform using gRPC (Protocol Buffers), Python, and Apache Spark:- ETL service to ingest raw data and store in different storage platforms (S3, GCS, Cassandra).- Spark cluster service providing platform-agnostic Spark cluster management and job submission (AWS ElasticMapReduce, Google Cloud Dataproc)- Data transformation service providing common operations (e.g. filter, join, split) and support for arbitrary transformations for preprocessing data for various machine learning problems such as time series forecasting and recommendation systems.

Berkeley, Ca

- Developed a Python interface connecting energy meters to vendor database via JSON POST requests.- Implemented customizable real-time update functionality to keep vendor database up to date.

University Of Michigan, Ann Arbor

- Developed a MATLAB automatic single unit recording and classification program using advanced signal processing and machine learning techniques.- Led code architecture design, algorithm development, and UI/UX implementation.- Improved cluster separation between units by implementing feature selection and sorting algorithms.- Optimized program to handle large datasets (~ 50 GB), dramatically reducing sorting time.

Berkeley, Ca

- Wrote a Python program to convert vendor energy data into open source sMAP format. Created Python framework supporting periodic updating of a sMAP server with new data.- Built a PostgreSQL database from ground up to consolidate energy data from different vendors in a streamlined format to facilitate data analysis.

Berkeley, Ca

- Optimized anode parameters to help improve Li-ion battery performance and durability.- Designed and conducted battery cycling experiments to gather data for performance analysis.

- Wrote scripts identifying inaccuracies in supercomputer generated air pollutant climate models.- Reduced time spent extracting emissions data from U.S. EPA databases by 80% by implementing macros.