Deep Learning Compilers

Deep Learning Compiler Verification

Deep Learning Compiler Verification

• Developed Spark Streaming pipeline to feature engineer internal and AWS data, improving efficiency for community detection algorithm

• Fixed data serialization issues of internal data product, saving development hours and increasing reliability

• Simplified pipeline by consolidating three Spark jobs into one, alleviating need for AWS resources

• Participated in Agile methodologies, including standup and two-week sprints, to receive constant feedback

• Developed data analysis tool in Perl to extract and organize metric data from memory compiler automation logs.

• Improved data collection accuracy by 20% by identifying critical bugs in the automation flow.

• Designed schema for MySQL database and wrote complex queries to interpret data and identify errors.

• Automated the generation of reports for performance analysis and status updates of batch jobs.

• Presented project results to panel of upper management and suggested future improvements.