David Bros

New at Essent.

- Deployed a GEO Redundant Kafka Cluster (8 nodes) in 4 POPs around the world, set up MM2 redundancy for real time data streaming. This cluster supports data from over 2500 customer planes, around 6.5 million documents every 5 minutes. (Kafka, Centos7) - Set up and maintained a 9 node Galera cluster and proxySQL for Observium (NMS System). This system is now in charge of monitoring over 2500 VMs in 4 different data centers in the world and provides critical status updates to hundreds of support engineers. (PostgreSQL, Observium, ProxySQL, MariaDB, Galera) - Worked on data parsing, processing, analyzing, and visualizing. Created and combined several internal and external data sources to enrich our client's datasets. (Ruby, Rails, ElasticSearch, Kibana, PostgreSQL, Kafka). - Designed and implemented a Data Model capable of supporting multiple Mobile Technologies and their metadata, regardless of source and data structure. This cut implementation times by over 4 hours per each Technology and standarized all sources. (PostgreSQL, Rails) - Designed a Spark cluster (Spark, PySpark, ZooKeeper) (6 nodes) which ingests data from data collectors located in the 3 world POPs. Provided near real time availability, 100% data consistency, greater ingest, processing and writing speed using map reduce and data micro batching streaming. (Spark, PySpark, ZooKeeper, Python, Ansible, Kafka). - Designed and developed a Geo Redundancy and Fail Over system for Delayed Jobs, this made all our data processors redundant and fail safe. It also allowed us to have 4 hosts with Delayed Jobs workers in 4 locations around the world. (Ruby, PostgreSQL, Rails) - Maintained and set up redundant, internet facing Kong Proxy, 4 VMs with custom Lua plugins and 2 custom PostgreSQL clusters (2 nodes) with versions 9 and 10 in a master-slave infrastructure. (Kong, Ansible, Lua, Keepalived, Centos 7, Yum).

- Developed 8+ data crawlers that track several public data sources, these fetched financial statements for hundreds of companies, making the Analysts have the latest reports on any company they were analyzed at the time. (MSQL, SSMS, Python, Docker, Selenium) - Developed a new Data Model for the Annual Insight's data platform, and took care of developing the necessary ETLs for that to happen. The model allowed for increased speeds when enriching data. (MSQL, SSMS, Python) - Created a PDF Parser with Python which read through crawled financial statements and saved the metrics read. These were then corrected or approved by analysts. The PDF Parser application saved about 4 hours of manual input for every analyst and every company they analyzed. (Python, OCR) - Designed, developed, and implemented an SDK for the company to use, the tools it included were: A generic Logging system that fed the monitoring dashboards with continuous statuses from every application, and, a Python Implementation of Azure AD for any of our applications to log in to Annual Insight's Key Vault and Secret Vault automatically. (Python, Azure SDK) - Designed and implemented testing and CI/CD with Azure DevOps. (Azure, Docker, Selenium) - Taught several Pythonic standards to the team in order to bring flexibility and reduce duplication and code debt.