Welcome! I'm Ming, a software engineer with a passion for building elegant and efficient solutions to complex problems. With over 8 years of experience, I specialize in backend development, cloud-native services, and creating systems that are both powerful and reliable.
A single LLM asked to evaluate your idea will almost always validate it. This is not a hallucination — it is a structural bias baked into how models are trained. Here is how adversarial multi-agent pipelines produce genuinely wider perspectives.
AI is only as precise as the intent you give it. Most AI development skips the hardest step, defining what you actually want, and pays for it in revision cycles.
AI code generation is a productivity win — until it quietly becomes a maintenance nightmare. Here is what is actually happening, and what to do about it.
Engineering solutions for ICE's post-acquisition mortgage data and workflow products.
Building software for end-to-end mortgage lifecycle management.
Developing automated valuation models used by lenders and appraisers nationwide.
AI-powered startup validator running a structurally adversarial multi-agent debate pipeline. Two agents, two opposing mandates — the bear reads the bull's full argument and rebuts specific claims, not generic risks.
ideology