Adrian Young@ap33zzi
2
Events attended
1
Submissions made
United States
2+ years of experience
About me
Founder & CEO of EcoPrint AI and a software engineering student figuring things out by building them. I’m interested in AI, infrastructure, measurement, attribution, developer tools, and sustainability—basically, anything happening behind the scenes that we usually don’t measure. I’m building EcoPrint around a simple idea: you can’t manage what you can’t measure. I love experimenting with new technology, tackling problems that seem a little weird at first, and turning ideas into working prototypes. I spend a lot of time exploring infrastructure measurement, energy and resource efficiency, and workload attribution. When I step out of my IDE, I’m a dad, a huge anime fan, and a COD Zombies expert. Honestly, it feels like I went back to college, stumbled upon a real problem, started building a solution that was way beyond my skill set, and months later looked up and realized I had become a software engineer. Still learning. Still building.
🤓 Latest Submissions
EcoPrint AI Transparency
Modern software development makes resource consumption difficult to see at the moment engineering decisions are made. Developers can see build times, test results, and application metrics, but often cannot connect a specific workload to the infrastructure signals observed during its execution, understand what changed between comparable runs, or distinguish attributable resource use from aggregate infrastructure activity. EcoPrint AI Transparency explores solving that visibility gap by bringing infrastructure measurement directly into the developer workflow through the CLI. A developer can wrap a build, test, or other workload with EcoPrint, which creates workload/run/attempt identity and process context, correlates the execution with SigSense infrastructure telemetry, and produces deterministic accounting and evidence. The prototype measures physical signals such as CPU/GPU/ANE power, total power, integrated energy, carbon, and water where supported. It also captures logical AI usage metadata such as provider, model, input/output/total tokens, and latency when the provider reports it. These domains remain explicitly separated: tokens are not energy, and attribution evidence is not the same as physical process-level isolation. EcoPrint adds baseline comparison, deterministic insights, process provenance, and optional AI interpretation through OpenAI or local Ollama. The AI operates downstream of deterministic measurement. Current measured facts remain authoritative, historical experience is contextual only, and unsupported AI explanations can be withheld when they conflict with available evidence. The intended users are software developers, platform/DevOps teams, and organizations building or operating compute-intensive and AI workloads. The broader vision is to move infrastructure and resource intelligence upstream—from something reviewed after deployment toward an observable engineering signal throughout ideation, architecture, development, testing, CI/CD, and production. Built with IBM Bob 2.0, EcoPrint demonstrates a working CLI-based approach to measuring, attributing, accounting for, and explaining infrastructure resource use while preserving clear measurement boundaries and evidence provenance.
27 Sep 2026
