SEO & AI search consultant
Karl Fung is an SEO and AI search consultant helping brands stay visible across search engines and AI answers. This site is where I publish field notes — build decisions, search experiments and things I learned doing the work.
Latest
Most automated game playtesting stops at generating bug reports or complaints. In Agency Lab, I wanted to close the loop: letting simulated player personas identify friction, formulate design proposals, and pass reviewed changes to a coding agent.
While building a semantic search and chatbot over classical Chinese literature, I ran retrieval benchmarks across embedding models and a three-mode ablation across eight LLMs. Here is what the success rates, failure modes, and tables revealed.
While building a marketing-agency sim, I wanted design questions grounded in what players already care about. This prototype collects Steam reviews of comparable games, turns their themes into personas, and carries the resulting suggestions through to reviewed code changes.
I am building a personal project diary where AI can propose articles from my repository work and the conversations I make available. A custom CMS lets me shape that writing process around my own thinking and ideas.