Tony Zhang

Designed and Implemented an auto-provisioning suggestion system to provide an appropriate number and type of AWS resources to meet both performance expectations and budget constraints.
• Profiled metrics for different iterative machine learning and data mining applications on Spark.
• Estimated performance and budget results based on the profiling metrics and AWS EC2 instance resources.
• Conducted comparisons and provided suggestions based on the estimation of performance and budget… Show more

Designed and Implemented an auto-provisioning suggestion system to provide an appropriate number and type of AWS resources to meet both performance expectations and budget constraints.
• Profiled metrics for different iterative machine learning and data mining applications on Spark.
• Estimated performance and budget results based on the profiling metrics and AWS EC2 instance resources.
• Conducted comparisons and provided suggestions based on the estimation of performance and budget results. Show less

Designed and Implemented a MapReduce Facility similar to Hadoop capable of dispatching parallel map and reduce processes across multiple hosts, as well as recovering from worker failure.
• Designed and implemented the basic MapReduce framework, including master and Slave, map and reduce, fault tolerance, scheduling and concurrency modules.
• Designed a pseudo-random data distribution function to look up address of slave storage devices without relying on a central metadata server.

Designed, developed and deployed a web service for Twitter data analysis that met the throughput, budget and query requirements of the client.
• Extracted, Transformed and Loaded process for six different kinds of client queries by extracting JSON Twitter dataset, transforming dataset into a specific format and loading the data into HBase/MySQL on AWS.
• Designed and built backend HBase and MySQL using AWS resources. Utilized MySQL Cluster and Elastic Load Balancer to meet the throughput… Show more

Designed, developed and deployed a web service for Twitter data analysis that met the throughput, budget and query requirements of the client.
• Extracted, Transformed and Loaded process for six different kinds of client queries by extracting JSON Twitter dataset, transforming dataset into a specific format and loading the data into HBase/MySQL on AWS.
• Designed and built backend HBase and MySQL using AWS resources. Utilized MySQL Cluster and Elastic Load Balancer to meet the throughput and budget requirement. Show less

Built a Cloud File System using file-system in user space (FUSE) based on different storage devices of SSD as well as Amazon s3 storage systems.
• Designed and implemented a hybrid Fuse file system to leverage the properties of SSDs and cloud storage by placing small data objects on SSD for fast retrieval and big data objects on the cloud storage.
• Designed and implemented block level deduplication of objects on the cloud storage to save space and network transfers, and a Cloud FS cache… Show more

Built a Cloud File System using file-system in user space (FUSE) based on different storage devices of SSD as well as Amazon s3 storage systems.
• Designed and implemented a hybrid Fuse file system to leverage the properties of SSDs and cloud storage by placing small data objects on SSD for fast retrieval and big data objects on the cloud storage.
• Designed and implemented block level deduplication of objects on the cloud storage to save space and network transfers, and a Cloud FS cache to improve the read/write operation efficiency by accessing data from cache rather than cloud storage systems
• Designed and implemented some snapshot related functions to benefit the consistency and recovery of file system, including create, delete, restore and list multiple snapshots Show less