Gregory Barron Email and Phone Number

Designed an AWS SQS strategy that managed the CARTO medical device data ingestion pipeline, leveraging AWS Lambda and S3 to improve downstream data quality by catching formatting and other data quality issuesDeveloped the database architecture for the CARTO, CARTONET, and Generator medical device data, and maintained the associated pipelines and models to accomplish data delivery objectives for multiple products owned by cross functional teamsCreated several testing methods ranging from device data verification to pipeline resilience, utilizing cloud-based systems such as the AWS CloudWatch API for near-live metrics handling and S3 for macroscopic verifications regarding file sizes and object counts for our large data ingestions into AWS RedshiftKept abreast of emerging AI/ML technologies by developing a proof of concept GPT3 model for generating SAPs, and also fine tuned a Llama2 model in AWS Sagemaker to create a comparison between the LLMsActed as a reliable source of information for anything data engineering related to support junior data engineers or other colleagues, and set up several training documents to provide a quick start for new hiresWrote and optimized complex SQL queries for AWS Redshift including materialized views utilizing UDFs for our business intelligence tables, and queries requiring multi-table joins to analyze data for TFLsWrote or modified several pipeline orchestrations in Python using AWS Step Functions, Lambda, Glue, SQS, S3, Redshift, EventBridge, CloudWatch, and created delivery mechanisms with AWS API Gateway, Lambda, and S3Created R-Shiny dashboards by following the SDLC including a transmission compliance analysis, patient profiles, and TFL statistical dashboardsPlanned out seminars hosted by AWS partners to encourage colleagues to use available resources, and to scope out possible solutions for J&J’s next generation products and data initiatives

Developed and deployed a PingID SSO enabled clinical study budgeting application for the biostatistics and data management team using Django and PostgreSQLPerformed administrative changes and updates to a series of EC2 machines that required tweaks to RAM usage to support the testing of high compute operationsCreated an R Shiny dashboard for a clinical study that leveraged Redshift SQL data to display summary statistics, and created formatted word document exports of the statistics using the ‘officer’ packageDesigned a cohort analysis using R Shiny to ensure compliance agreement for all applicable studies, which allowed for user entered criteria and time framesDetermined the root cause of a data duplication/multiplication issue concerning Wrike data ingestion, and modified the lambda function containing the errant upsert methodCreated lambda functions to support a reporting pipeline, ending with a formatted HTML file saved to S3Put together the budget versus actuals dashboard using Tableau by joining together the budget application dataset to the Wrike data through a complex SQL query

Led a team of developers by designing interfaces and API endpoints, using Jira with agile methodologiesCreated a highly available, resilient thirteen node Elasticsearch using Ansible and DockerDesigned a Redis caching strategy for company dashboards to enable performant large data analysis to effectively compete with technologies like Kibana and TableauDeveloped a system similar to AWS Lambda which walked through a large data CIFS share, and used any changes in modified date metadata as events, which caused a trigger to take the file path and perform our ETL to prepare the data for analysis through the use of Redis bots, Python, and ElasticSearchKept abreast with new technologies and applied findings to enhance our business use casesTook 100% ownership of the entire platform, transforming it from MVP to validated productWorked with clients to deploy the PaaS behind their firewalls through the use of VPNs, Linux jump servers, and network tunnelsDesigned and implemented the variable connector system, capable of ingesting data into Elasticsearch at scale from any source, including third party APIs, git repositories, databases, and file systems after normalization and enrichment to support ETL in a similar fashion as AWS GlueDeveloped, maintained, and tested Django APIs and qcluster job services for our flagship PaaSCreated the fully automated deployment on Linux for the PaaS using docker, ansible, and bash, including a development deployment that enabled SSH access to mounted docker volumes for use with the remote explorer in VSCodePerformed ETL on two petabytes of data that were spread across NFS shares deployed on AWS EC2 instances, and completed the analysis in one month using a combination of concurrency, multiple jobs on Django’s qcluster, and several high performance EC2 instances

Developed an enhanced poll weighting algorithm and the associated data models guaranteed toproduce live results in less than two seconds for big data (tested on upwards of 100 million pollresponses at once)Developed custom testing by modifying kertosis/skewness of randomized responses, and usingcontinuous distributions to generate the randomized responses (uniform, normal, cauchy, and exponential)Leveraged available big data APIs and FTP servers concerning census data, and otherwisepolitical data to inform users of upcoming elections and provide pertinent information about thecandidatesProduced statistical reports demonstrating user activity and feature usage to drive sales anddevelopment decisionsCreated localized data feeds by via census data, available local news APIs, and available pastpublic voting information

Connected code built in C, perl, and matlab to produce a short simulation of a blinking eye usingseveral terabytes of dataDeveloped foundational gradient descent or other optimization algorithms in numerical linearalgebra that served as the backbone of several machine learning methodsUsing python, R, or matlab to solve stochastic processing problems in topics including markovchains, birth-death processes, Bernoulli processes, and random walksStudied stochastic modeling and critical path analysis for project management