Hi,

I like making complicated things simple. I usually start by going all the way down — breaking a problem apart, looking at what is happening at every level, from the raw fundamentals to the thing sitting in front of us today. Then I build it back up into something that works. Automation is usually the answer. AI is too, when it actually makes sense. I use AI where it genuinely makes a system better, and never just because I can.

I have a fairly diverse taste in things — design, music, philosophy, art, media, engineering. I’m quite an artist, and a very serious engineer. I’ve always believed that great engineers make good artists, and great artists are usually good engineers. I’m most comfortable when something is complicated, slightly perplexing, and still has no obvious shape. I like giving it that shape, putting it into structure, and making the whole thing work smoothly. That’s quite me.


So far

Animoart Productions Foundational Chapter
AI Automation Engineer
May 2024 — June 2026
Learned how creative production actually works end to end — building reliable, day-to-day automation pipelines across shoots, edits, and campaigns.

Animoart was where I learned how creative production actually works end to end. I got exposure to the full cycle — from client enquiry and planning through shoots, editing, revisions, and final delivery — and to the operational side of events and productions as clients regularly brought us in for shoots and on-ground execution.

Working that closely with production taught me something that shaped how I build automation: a small system that does one job reliably is usually more valuable than an ambitious system that tries to do everything and keeps breaking. The best automation was often the simplest piece of infrastructure that a team could depend on every day.

I contributed heavily to the agency's adoption of generative AI, building client-specific workflows and campaign-specific pipelines that helped graphic designers and video editors create posts, reels, mockups, and other creative assets much faster. In many cases, the generated work was close enough to be approved with only small changes; when it wasn't, it still gave the team a strong starting point or creative direction instead of a blank canvas. I built and maintained roughly 100 workflows across production, media, research, feedback, campaigns, web, and business development, and also trained designers and editors to incorporate these tools into their existing work.

The Brand Research, RAW Culling, Video Indexing, Feedback Resolver, and Asset Handoff systems on this site are public-safe recreations of the kinds of problems that actually came through this work.


FUTURE TRAJECTORY · EXPLORATIONS

Where I’m Going

My interests might look a little scattered from the outside. They really aren’t. They’re mostly different parts of the same line — understanding how things work, figuring out how they can be made better, and then finding ways to make them real through code.

01 — Creative Systems

Programmatic Video

I’m interested in how creative media gets made when software becomes part of the creative process itself. Not just generation, but hybrid systems where image and video models work alongside code and coding agents to build media that is more structured, controllable and expressive.

02 — Theoretical Foundations

AI & Computational Mathematics

I was always good at mathematics. It was one of those things that felt intuitive to me, and machine learning came naturally for much the same reason. I’m academically inclined towards the raw side of AI — the mathematics, the logic, the way a model learns, reasons and changes. Lately, I’ve been especially interested in self-improving AI: what the mathematics and logic behind it actually look like, how a system can evaluate itself, find where it went wrong, and use that information to become better.

My professional direction is Applied AI. That is where I can take that understanding and put AI into actual use and perspective — into systems, workflows and products where it has to work in context, not just in theory.

03 — Engineering & Operations

Applied & Self-Improving AI Systems

This is the part I’m especially interested in building. AI systems that do not simply execute a workflow once, but can handle ambiguity, detect when something has gone wrong, recover, and improve from what happened. I’m interested in the engineering around that — the workflows, pipelines, evaluations and feedback loops that make AI genuinely useful and increasingly reliable in real environments.