Matching engine delegation
You're implementing a novel order-matching algorithm for a low-latency trading engine in C++: price-time priority with a custom pro-rata allocation rule that only your firm uses, and it must reconcile bit-for-bit with an existing reference implementation. A teammate suggests letting an AI agent write the whole matching loop from your one-paragraph description. Walk through whether you'd delegate this, and if not, what failure modes make an AI assistant the wrong primary tool here.
allocate_pro_rata(order_sizes: list[int], incoming_qty: int) → list[int][[100,200,300],60]out[10,20,30][[100,100,100],10]out[4,3,3][[10,20],100]out[10,20]Treat the AI’s output as a draft to verify, not an answer to trust. Name the specific flaw and the input that triggers it, say how you’d catch it (tests, edge cases, reading critically), and how you’d re-prompt or decompose to get it right.
Vibe & agentic: describe the solution in plain language (or narrate it) and the coach grades your approach.