AI & Convergence
Where everything comes together
Practical AI work in daily teams and hands-on building with Claude Code, OpenAI Codex, and manual coding. The place where visual, audio, and systems experience converges.
The Long Arc
Some paths are only visible in retrospect. Growing up with Star Trek, Star Wars, and The Hitchhiker’s Guide to the Galaxy built an imaginative foundation — a sense that technology could be a creative partner, not just a tool. Decades of working across visual, audio, and digital media felt like separate threads. With AI, they’re finally weaving together.
This isn’t nostalgia. It’s the recognition that a multimedia background — understanding how images work, how sound communicates, how systems are designed, how people collaborate — is precisely what makes AI work meaningful. Multimodal AI doesn’t just process text, images, and audio; it requires people who understand these mediums deeply enough to guide, evaluate, and build with them.
AI Enablement in Practice
In my company role, I help teams move from AI curiosity to practical AI fluency. The most effective way to drive adoption isn’t presentations about capabilities — it’s showing workflows that solve real problems.
This means demonstrating prompting techniques, helping colleagues integrate AI into their specific workflows, curating tools and methods, and building a culture where experimenting is normal and useful. Everyone’s job is different, so the AI patterns that matter to a marketer differ from those that matter to a developer.
The work is mostly about lowering barriers, building confidence, and listening more than talking.
Building with AI
Signal Lens
This site is a meta-portfolio piece: a website about AI, built with AI. I built it collaboratively with Claude Code and OpenAI Codex, with plenty of hand coding in between. In practice, that means architecture decisions through conversation, multi-file implementation with agent help, and direct manual edits where precision matters.
I also use Grok selectively for real-time news context, mainly as a research input when timeliness matters.
The tech stack (Astro 5, React 19, Three.js, TypeScript, Tailwind CSS v4, Cloudflare Pages) reflects deliberate choices. The interactive concept explainers — attention mechanisms, tokenization, embeddings — are portfolio pieces in themselves, demonstrating both technical depth and visual communication.
Signal Deck
A privacy-first AI knowledge workspace running inference locally through Ollama. No cloud calls, no data leaving the device. It demonstrates that local AI can deliver a smooth, practical experience — summarization, reasoning, OCR, multi-step pipelines — while respecting privacy. Building it reinforced understanding of model selection trade-offs, inference optimization, and designing AI tools that put the user first.
The Bigger Picture
AI is democratizing creative and technical work. More people can turn ideas into usable output with less friction, across media and software.
The human element still matters most. AI doesn’t replace judgment, taste, or collaborative instincts built through experience. It amplifies them when used well.