Cliff Wells Email and Phone Number

During my tenure at PDL, I was tasked with redesigning an internal Django application used by sales engineers for managing customer accounts. The original application was essentially bare HTML reports. I redesigned the UI using jQuery and Bootstrap 5, making it not only more attractive and user-friendly, but also improving performance by adding pagination and caching to computationally-intense reports, and adding a multitude of new features.I also rewrote an existing ETL pipeline that was too slow and memory-intensive. This pipeline extracted data via S3 Select, did some processing, and generated reports for customers. Some reports that had taken up to 9 hours to complete were now able to complete in under an hour. I changed the architecture from a simple thread pool to a series of producer/consumer queues demarcated by their function and allocated processes to these queues based on their bottlenecks (CPU, network, memory).

Startup developing telemetry-as-code product for monitoring uptime/performance of arbitrary internet services.Rather than utilize a traditional API/GUI approach for managing checks, I implemented a YAML/JSON configuration to allow for deployment of checks along with other IaC. During the year I worked on this, I developed the entire stack, from Terraform/Serverless deployments for various portions of the stack, a YAML-to-Python compiler that was the heart of the system, as well as utilizing Celery/RabbitMQ and systemd to manage the system's backplane. Project put on hold due to pandemic.

During my time at FocusVision, I helped manage a bit over a hundred colocated servers as well as AWS instances and services. I also helped manage MySQL and PostgreSQL databases, but mostly wrote thousands of lines of Puppet code to bring these systems, which were mostly bespoke, under automation. This company had acquired multiple products written in as many languages (Python, Ruby, Elixir, etc), each using different databases (MySQL, PostgreSQL) and each product had come with its own hosting platform and varying levels of automation using different tools, from bespoke systems, to custom Bash scripts, Ansible, and Puppet. My team decided to use Puppet to unify management of these systems, so I spent the bulk of my time there writing Puppet manifests to manage all aspects of these systems. This involved significant forensics to ensure we captured the full functionality of these systems in Puppet manifests.I also developed a custom Puppet module in Ruby for managing Pingdom checks to improve the organization's monitoring coverage. When I started there were only a handful of hand-managed checks on key systems. Using this module, I was able to automatically deploy Pingdom checks and achieve 100% coverage in only a few hours.Aside from technical work, I also worked with various product managers to help plan how applications would be deployed, helping set timelines and expectations, as well as offering advice on technical decisions.

Member of team that managed the application's platform and deployments using CloudFormation, Chef, and other tools. Responded to incidents, and ensured the platform was robust and secure. I was also tasked with resolving the single security issue the company faced during my tenure, investigating a compromised system to discover how the breach occurred, estimate the extent of the damage, and rebuilding and securing a replacement system.I also worked with the front-end team to help them automate the deployment of the company's corporate website, and developed a process for reducing utilization of AWS resources by our QA team.

Member of the infrastructure team that was tasked with bringing a patchwork of CFEngine-managed and bespoke systems under Puppet management. Wrote Puppet manifests that managed every aspect of system lifecycle, from deployment on bare-metal, including configuring iSCSI/NFS resources on NetApp cluster, database connections. This was fairly involved as the company had many legacy systems that were on outdated Linux distros and whose purpose may not have been entirely clear anymore. As such a significant part of the work involved forensics to decipher the requirements that needed to be managed by Puppet, while simultaneously not interrupting services.I also was the lead on a team that developed a report queuing system that solved a longstanding issue with some legacy reporting systems becoming overloaded. This project was part of the annual "Innovation Days" contest and ours was the winning entry, solving an issue the CEO had "waited years" for.

Combination web hosting / web development company focused on providing internet services to small-to-medium businesses in Portland area. It was during this time that I also created and hosted the original Nginx wiki.I ran this business from top-to-bottom, building a colocated hardware platform starting with shared hosting for customers that I was building websites and web applications for.Eventually I upgraded to a cluster of servers, using OpenVZ for virtualization, and using a fibre channel backplane to a set of SCSI arrays for storage. As the cloud became dominant, I decided to exit this business and sold the platform to a local business.

I was initially hired by this company to take over development of a web application that allowed customers to manage call forwarding and caller ID. Having brought this project to the maintenance stage, I began helping them develop an IVR system using FreeSWITCH, and expanded from there into setting up PBX systems in colocation facilities across several states. I helped debugging SS7 and SIP issues, and eventually was helping them develop their telephone network from the bottom-up: splitting DS3 drops into T1 lines, writing software to manage call routing, and integrating everything into the front-end application I had initially been hired to write.

During my tenure here, I developed web applications in Python and PHP. I also helped managing the hosting platform, managing services such as Apache and Sendmail.This company specialized in PostgreSQL support and part of my job was to assist customers with PostgreSQL-specific problems, including optimizing complex queries, migration of data, and other DBA-related tasks.

This company had a proprietary product used for physical access control (e.g. card/badge readers) that was written in C and ran on SCO Unix. I not only added features and and fixed bugs in the management application, but also installed the OS and application on computers that would be delivered to remote locations, helped customers solve problems remotely.Given that our customers were primarily high-security facilities such as military bases, universities, police stations, etc, delivering a reliable product and quickly dealing with issues was paramount. I also recommended and started implementation of a next-generation version of the platform that would move the company away from a proprietary database to PostgreSQL, and move the front-end application from a bespoke C console application to one written in Python and utilizing curses for customers that needed a console UI, as well as a graphical version using wxPython for customers who preferred this.I further worked with our hardware engineers to co-develop a new USB/ethernet interface in preference of the previous RS-232 communications protocol.