Christian Santiago

My work has spanned both ends of the scale, and each end teaches something the other cannot. Enterprise infrastructure, where a change to a shared system has a blast radius and you learn to respect it. Small and mid-sized businesses, where there is no layer between you and the client who needs the answer, so you gather the requirement, build the pipeline, and present the result yourself.

I spent four years in an enterprise environment on infrastructure monitoring and incident response, and the data engineering around it: thresholds and event-driven alert routing on the detection side, telemetry pipelines and dimensional models on the side that turned resolved incidents into failure patterns. All of it through an on-premise to hybrid AWS migration. Since then, systems integration, ETL pipelines, BI reporting for leadership, and freelance application development.

The through line in my work is system design. Good architecture has less to do with knowing services than with asking what tradeoffs you are making with a decision, and being able to answer. Reliability against cost. Simplicity against flexibility. Shipping now against unwinding it later. There is rarely a correct answer, only one that fits the requirement in front of you, so I reach for the smallest thing that meets it rather than the largest thing that impresses. I treat those decisions as revisable rather than permanent, and let telemetry rather than my own opinion tell me when one needs revisiting. The habit came from presenting technical decisions to leadership and non-technical stakeholders, who were never going to accept "because it is best practice" as an answer.

Lately I have been pointing that at generative AI and machine learning workloads, in Go, Python, and Terraform on AWS. MLOps is the direction I am headed. I am an AWS Certified Developer, studied computer science at Vanderbilt University and cloud engineering through the AWS Cloud Institute.

Most of what I do outside work comes down to getting people in the same room. That looks like cooking dinner for a friend group, or pulling friends who live in different cities onto a call to play through whatever strange indie game somebody turned up on Steam. The activity is mostly a pretext. The part I actually care about is that everyone showed up.

I cook a lot, and I recently restored my first cast-iron skillet, which took considerably longer than I expected. Mostly sanding rust off and putting thin coats of oil on it in the oven, over and over, until it stopped being orange. Worth it, though. Cast iron holds heat the way a thin pan cannot, so food hits the surface and actually sears the way you see in cooking videos.

I am in Nashville, so on a given week you might find me at a community event at my local bodybuilding gym, at a developer meetup at Bitcoin Park, or riding my road bike through a state park with no particular destination in mind.

I like learning for its own sake, and I have never really stopped. At the moment that means grinding through the long path of Math Academy's Mathematics for Machine Learning track, keeping up with deeplearning.ai, and working through DataCamp's AI and machine learning courses, then going and building something with whatever I just picked up. That last part is the honest explanation for most of the projects on this page. They started as things I wanted to understand rather than things I needed to ship.