Questions an interviewer could realistically ask about this project, drawn from
../learning/. This project is unusually good interview material because it’s an
agentic system that takes irreversible actions on your behalf — which is exactly the design
problem every company building with LLMs is currently arguing about internally.
How to use this
- Answer out loud, from memory, before reading the hint.
- Strong answer covers = the checklist a good answer hits, not a script.
- [Beginner] / [Intermediate] / [Advanced] = how much project context the question assumes.
- 🔥 = most likely to be asked.
Files
| File | Covers |
|---|---|
| 01-pipeline-and-system-design.md | Stage costs, idempotency, dedupe keys, autonomy tiering, degradation, retries |
| 02-ai-and-llm-engineering.md | BYOK, provider abstraction, AI SDK v7, structured output, batching, verify() |
| 03-judgement-and-story.md | The ethics question you will be asked, trade-offs, shell/tooling habits, behavioural |
Before anything else: the 60-second pitch
A pipeline that discovers jobs across multiple sources, filters and scores them, tailors a
résumé per job, and then either submits or queues for review — designed to run unattended at 3am.
It’s bring-your-own-key: the config names an environment variable, never holds a secret, so it
works against Claude, OpenAI, Google, OpenRouter or a local Ollama with a one-line change. The
two design ideas I’d defend hardest: stages are ordered cheapest-first so free local filtering
shrinks what the expensive model call ever sees, and anything the model generates gets
mechanically verified against my actual profile before it can reach an application — because
a prompt saying “don’t invent skills” is a request, and a fabricated line on a résumé is
something I’d have to defend in an interview.
The last clause is the whole pitch. Lead with the guardrail, not the automation.
The one question to over-prepare
“You built a bot that applies to jobs for you. Isn’t that spam?”
You will get this. It is not hostile — it’s a test of whether you thought about it. The answer
lives in 03-judgement-and-story.md, Q1. Do not improvise it.
Software engineering
Infrastructure & security
Computer science
Hardware & engineering