About

DevOps and Cloud Engineer with 20+ years of platform engineering: CI/CD pipelines, containerized applications, infrastructure as code, and production monitoring on Linux with Ansible, Terraform, Docker, Kubernetes, Python and Bash across AWS, Azure and GCP. Containerized Jabil's iFactory platform onto Kubernetes across ~112 plants, automated Raymond James releases with Ansible and Jenkins, and ran Sprint's 50-server billing mediation estate at 99.999% uptime for 14 years as the on-call Tier 3 owner: 70% faster deployments, 60% fewer deployment errors, 50% lower alarm MTTR, $2M saved through custom routing software. Production Linux since 1996.

30
years of production Linux, 1996 to now
3
public clouds automated: AWS, Azure, GCP
14
years carrying the Tier 3 on-call pager for Sprint billing mediation
6
OS distributions package corkscrew: Debian, Ubuntu, Red Hat, CentOS, FreeBSD, Cygwin

Experience

Same bullets as the PDF, same order; nothing embellished for the web. Twenty-five years of somebody else's production, from dial-up ISPs to a global manufacturing platform.

Founder & CEO / DevOps Engineer VimOps, Kansas City metro (remote)

  • Cut client deployment times 70% by automating infrastructure with Ansible, Terraform, Docker and CI/CD (GitLab CI, GitHub Actions) across AWS, Azure and GCP, improving multi-cloud reliability.
  • Reduced build failures by over 50% by architecting scalable CI/CD pipelines in GitHub Actions and Jenkins.
  • Maintained 99.9% client application uptime with automated backup, disaster recovery and server provisioning workflows in Ansible and Docker.
  • Reduced deployment-related service disruptions 90% by refactoring a client's legacy monolith into containerized microservices.
  • Improved system reliability 80% by deploying and operating highly available Kubernetes clusters for client workloads.
  • Designed cloud migration strategy for 30+ clients, reducing infrastructure costs by an average of 75%.

Senior Infrastructure Automation Engineer Jabil, St. Petersburg FL

  • Reduced deployment time for critical iFactory infrastructure by 50% by building CI/CD build and release pipelines in Azure DevOps and Octopus Deploy with automated testing gates.
  • Cut deployment errors 60% with Terraform infrastructure as code and centralized ELK logging, making builds repeatable.
  • Cut new-developer onboarding time 30% by containerizing all iFactory applications with Docker, breaking siloed apps into microservices, and deploying them on Kubernetes across ~112 manufacturing plants.
  • Reduced manual configuration time 70% with a comprehensive Ansible playbook library for deploying, configuring and maintaining on-premises and cloud resources.
  • Drove DevOps adoption for iFactory across ~112 plants: canary, A/B and blue-green releases in Azure DevOps and Octopus Deploy, plus CI, automated testing, alerting and documentation standards replacing manual releases.

System Administrator (contract), Capital Markets Raymond James, St. Petersburg FL

  • Cut release deployment times 50% with Ansible playbooks and Jenkins pipelines that automate releases across test and production environments with consistent configuration.
  • Anchored Linux SME and Tier 3 application administration for the Capital Markets team's GBST Syn back-office trade-processing suite in a highly regulated financial environment.

Product Manager & Software Engineer codesigned, Atlanta GA (remote)

  • Designed and maintained on-premises and Azure infrastructure (site-to-site VPN, routing) for 20 large enterprise customers.

Software Engineer, Accounting Operations / Billing Mediation Sprint (now T-Mobile), Overland Park KS

  • Decreased deployment errors 30% and sped releases 40% with a production change-management system built on Ansible and GitLab CI.
  • Cut time-to-resolution for actionable alarms 50% by deploying the ELK stack and HP OpenView for observability and alerting across multiple mediation platforms.
  • Anchored Tier 3 support, root cause analysis and on-call escalation for 14 years across every Sprint billing mediation platform, resolving critical issues with a 98% success rate and reducing service downtime 20%.
  • Saved $2 million annually versus a vendor solution by designing and leading development of an automated EDR routing application (Perl, Python, shell) that distributed switch usage records to rating, billing, fraud and analytics systems.
  • Sustained 99.999% uptime (about 5 minutes/year) for 14 years across the 50-server bare-metal and virtualized RHEL/CentOS/Ubuntu mediation estate through disciplined patching, vulnerability scanning, hardware selection and vendor acceptance testing.

Founder CyberRazor, part-time alongside Sprint

  • Built Fasttrack, a Ruby on Rails web application for standard celeration charting used by behavior analysts, and trained educators on it; analyzed HCRIS healthcare cost data for tax firms; built small-business sites.

Systems Administrator and Network Engineer various ISPs and consultancies

  • Ran Unix systems and networks for six dial-up-era ISPs and consultancies (Advertisnet, ISPN, Century Computer Consultants, Interdirect/SupplyTrak, Cephas, Chemtrol): a 12-server RADIUS, SMTP/POP, FTP/web and billing-database stack serving 3,000 subscribers. Migrated Advertisnet from Windows NT to Linux/BSDi at sixteen.

Skills

drwxr-xr-x
IaC and config
Terraform, Ansible (playbooks, roles, YAML, Jinja2), CloudFormation, Puppet, Chef, Salt
drwxr-xr-x
CI/CD
Azure DevOps, GitHub Actions, GitLab CI/CD, Jenkins, Octopus Deploy, AWS CodePipeline, Git, pull-request review and branching/tagging strategy; canary, A/B and blue-green deployment
drwxr-xr-x
Cloud
AWS (EC2, S3, RDS, VPC), Azure (VMs, Storage, Virtual Networking, Azure DevOps), Google Cloud Platform, OpenStack, hybrid cloud, migration and cost optimization
drwxr-xr-x
Containers
Docker, Kubernetes, microservices and service mesh, HashiCorp Nomad and Consul, LXC, Podman, KVM, VMware ESX/vCenter/Horizon, VDI
drwxr-xr-x
Observability
ELK (Elasticsearch, Logstash, Kibana), Splunk, Dynatrace, Prometheus, Grafana, centralized logging, alerting and dashboards, SLOs/SLIs, on-call incident response, postmortems and RCA, SRE practices, ServiceNow/AIOps
drwxr-xr-x
Linux and OS
RHEL, CentOS, Debian, Ubuntu, HP-UX, Solaris, FreeBSD, Red Hat Satellite, Windows Server, Active Directory / IAM
drwxr-xr-x
Languages
Python, Bash / ksh / zsh, PowerShell, Perl, Ruby, PHP, SQL, C / C++ / C#, Java, JavaScript / TypeScript
drwxr-xr-x
Databases
Oracle, MySQL, PostgreSQL, SQLite, Microsoft SQL Server
drwxr-xr-x
Operations
on-call and Tier 2/3 escalation, SOPs and runbooks, release and change management, capacity planning, backup and disaster recovery, SLA adherence, ITIL, Agile / Scrum, Jira, Confluence
drwxr-xr-x
Network and security
IPv4/IPv6, routing, firewalls, load balancers and reverse proxies (nginx, Apache), Cisco IOS, Wireshark / libpcap, POSIX sockets, DNS, SSH, RADIUS, site-to-site VPN, VoIP / SIP / Asterisk, vulnerability scanning and patching

Field notes

Short posts on blog.patpadgett.com about the work behind the bullets: what the number actually meant, what broke, what I'd insist on now.

  1. Twenty-five years of corkscrew. All 291 damned lines.

    26-year-old C under GCC static analysis, ASan and a hostile mock proxy. What it found, and what I'm fixing.

  2. Never post twice. Yes, even when I run the damned thing twice.

    A cron publisher with a SQLite ledger and three statuses, because sometimes you genuinely don't know whether the POST landed.

  3. Two hundred and fifty percent. Yes, folks, it was Perl.

    Five parallel processor chains, round-robin by switch, strict sequence within each. The hard part was keeping the end of a call behind its beginning.

  4. NT to Linux at sixteen. Yes, I still had homework.

    Migrating an ISP's service stack to Linux/BSDi while in high school. The servers were the easy part; somebody had to answer the phone.

  5. Fourteen years at five nines. Yes, we patched them.

    Fifty servers, 99.999%, and how you get there by touching the machines, not by being afraid of them.

  6. Alarm fatigue: your monitoring system found the caps lock key

    ELK and HP OpenView cut resolution time for actionable alarms in half. Tell the operator what to do; making noise is the easy part.

  7. Runbooks for a smaller team. Write down the damned part you know.

    Twelve people to four without losing coverage. At 3 a.m., 'verify processing is normal' is not an instruction.

  8. Change management before DevOps: did billing get the damned records?

    Ansible and GitLab CI for production change control. A green run proves tasks completed; it doesn't prove the records arrived.

  9. Containers across 112 plants: green doesn't mean go

    Docker, Kubernetes, Azure DevOps and Octopus Deploy for a global factory platform. My confidence is not a test gate.

  10. Your installer finished. Is the damned system working?

    Splunk, Dynatrace, Ansible and Jenkins inside a financial change window. Faster deploys buy time to verify; don't spend it.

  11. Idempotency: can I run the damned thing again?

    What Ansible, shell scripts and ledgers each promise when you lose the terminal and run it twice.

Recommendations

Verbatim. The first two are from Sprint leadership, the third from the codesigned founder; the last is a Linux Magazine review of corkscrew.

Message from Kathryn Walker (Chief Network Officer, Sprint):

Patrick, your display of tenacity, creativity, and dedication through your perseverance and hard work for project D3884 delivered exemplary results! You beat the aggressive schedule, total costs were less than budgeted and service impacts to current customers were avoided.

EOF

Message from Renee Keffer (Director Network Operations, Sprint):

Pat, thanks for the quick response for implementing an alarming solution for the ProDiff files. You have been instrumental in maintaining this project, and your documentation has helped others gain an understanding of the processes involved. Your ingenuity is well known among many here at Sprint and is not going unnoticed.

EOF

Message from Jake Weaver (Founder & CEO, codesigned):

Along with his impressive technical skill set, Pat has been great to work with. He is a positive, upbeat personality that complements any team he is a part of.

EOF

Message from Charly Kühnast (Linux Magazine, September 2014):

The right tool for tunneling SSH through a proxy like Squid, Gauntlet, CacheFlow, Junkbuster, or Apache mod_proxy is Corkscrew. Why? Because the name corkscrew is well chosen, and because the program author introduces himself like this: 'My name is Pat Padgett. I'm a dork.' In my experience, such laconic people are great programmers.

EOF

CORKSCREW(1)User CommandsCORKSCREW(1)

Corkscrew and the 2026 tools

NAME

corkscrew - tunnel OpenSSH through HTTP and HTTPS proxies

SYNOPSIS

ProxyCommand corkscrew proxy.example.com 8080 %h %p

DESCRIPTION

C program that tunnels OpenSSH through HTTP and HTTPS proxies; 194 GitHub stars, 36 forks, GPL-2.0. Packaged in Debian, Ubuntu, Red Hat, CentOS, FreeBSD and Cygwin; reviewed in Linux Magazine (Issue 166, 2014) and 2600: The Hacker’s Quarterly (Vol. 19, Issue 2, Summer 2002); has its own Wikipedia article. Maintained 2000 to present.

CI

2026 modernization: GitHub Actions CI with regression tests and ASan/UBSan sanitizer builds, updated Autotools checks, and a documented security model (SECURITY.md). Reviews and merges community pull requests; supports distribution packagers.

APROPOS

Three more tools shipped in 2026, all Python 3 standard library only, MIT, with tests:

undo (1)
the shell undo button: journal-of-inverses wrappers for mv, cp, rm, chmod, chown, ln, mkdir and rmdir that work out the reverse operation before the real command runs, trash-protect deletes and replay the inverse on demand. 18 unit tests plus a shell suite.
whom (1)
who should I ask about this file, and are they awake? Ranks contributors from git history with a recency-weighted score, infers each person’s local time from their commits and flags who is likely asleep. On PyPI as whom-cli.
huh (1)
what is this string? Paste an epoch, JWT, cron line, chmod mode, exit code, CIDR, UUID, base64 or hash and get the plausible readings ranked, each with the Unix one-liner that reproduces it. 56 detectors, 110 tests in GitHub Actions, on PyPI as whatsthis.

SEE ALSO

github.com/patpadgett/corkscrew · Corkscrew (program) on Wikipedia · all repositories

Education

Cert
Certificate, Graphic Arts
School
Lake Career & Technical Center, Camdenton, MO
Years
1996–1998
Langs
English (native); Indonesian (professional working proficiency).
Study
Self-directed DevOps and infrastructure-as-code study, 2017–2018, which led directly to the Jabil role.
Plan:
2nd place, State of Missouri, VICA (Vocational Industrial Clubs of America) competition in offset lithography, running AB Dick and Ryobi presses.
Learned to keep a press registered before learning to keep a deploy pipeline registered. Same instinct.

Contact

DevOps, cloud, platform or SRE roles; production support for systems that cannot go down. Remote, hybrid, contract or full-time. Tampa Bay based.

pat@patpadgett.com
+1 816-601-5983

Or grab the file and skip the form:

-rw-r--r--47KPatrick_Padgett_Resume.pdf -rw-r--r--40KPatrick_Padgett_Resume.docx -rw-r--r--7KPatrick_Padgett_Resume.md
To: pat@patpadgett.comSubject: job

Fill this in and it opens a ready-to-send draft in your mail app — or just email pat@patpadgett.com.

Add your name so the reply has a greeting.
A working email address, so the reply can reach you.
A line or two about the role is enough.