About

I came into tech in 2011 through Workforce Opportunity Services, placed at Prudential Financial while I was still serving in the NJ Army National Guard, and taught myself to code on the job. By 2013 I was the Remote Access Analyst: Tier 3 help desk work with a side project generating QR codes for MFA enrollment. That side project ran for eight years and produced over 300,000 activation codes. It also settled which job I wanted: the engineering one, not the support one.

From there I moved up through Remote Access SRE and Senior Remote Access SRE at Prudential, where I built the MFA self-service portal used more than 18,000 times in six languages, automated over 200,000 administrative actions, and kept the VPN running for 60,000 corporate users through the early months of COVID. Most of that work was figuring out how to make safety-critical infrastructure quiet: how to remove failure modes rather than paper over them.

My last year there, from July 2021, was as a Cloud Engineer in Site Reliability Engineering: hands-on AWS and Azure ownership at Fortune 100 scale, building the monitoring flow that alerted leadership channels to global cloud outages in real time and partnering with application teams to unblock deployments between on-prem and the cloud. It was a deliberate pivot, and it is the year that made the next move possible.

In June 2022 I joined Amazon, where I built a monitoring lifecycle platform that onboarded more than 2,750 application stages and brought 3 million monitors under standardized management, and hired and mentored eight new-to-industry engineers in New York. In August 2024 I moved into my current role as a Senior Software Engineer on a platform team, architecting a multi-region workflow orchestration control plane whose guardrails have run 500,000 safety checks, validating every change before it lands. Since then I have mentored engineers on three continents. The lesson carried over from Prudential intact: the platforms worth building are the ones that stay quiet while everything they watch does not.

Outside work I run the same convictions as an engineering team of one: one engineer, a fleet of AI coding agents, and a spec-driven method where every feature starts as a written spec and nothing merges until it passes the same gate the pipeline runs. The grading is public. Caskey Engineering carries the case studies, a lesson series that teaches the setup one file at a time, the numbers posts that count what it produces and what it costs, and backtests that report what did not work. The builds include Ballast, an open-source RAG system engineered to refuse when it should; SpecSelf, a spec-driven operating system for a life whose agents read but never write; a C++ options engine tuned to 215 million prices a second; and AI finance tools where the model narrates numbers it is never allowed to derive. The selected builds are on the work page; the full writeups live on Caskey Engineering.

Along the way I ran a New Jersey WordPress shop called Caskey Coding from 2015 to 2018 (ten sites for small businesses and community organizations), and served six years as a Military Police Officer in the 50th Brigade of the NJ Army National Guard, including a 2012 Superstorm Sandy activation where I safeguarded the distribution of 241,000 gallons of FEMA fuel for nearly a month.