generate 10 questions for this position be very specific in these question and refer concrete details from the file analyze the file and also generate questions that address gaps in the file compared to the role description make sure that the file content is relevant for the role description. if not the case, simply reply with "candidate irrelevant" role description role: ai developer - client: company from germany - product: process automation - experience: proven experience in building llm apps, orchestration, prompt templates, chains, llm apis, prompt design & optimization - tasks: - refine llm api calls - refine prompts for data extraction - extract data from pdfs (via text layer and ocr) - orchestrate api calls, build prompt sequences & chains - integrate with microsoft outlook, sql, rest api connectors - build an installable package to automate a workflow - llms: gpt, claude, llama, mistral - languages: python, sql - libraries: openai, anthropic, ragas - location: 100% remote --- review your questions one by one by printing them out, then validate they are consistent with the file contents and the role make sure that questions are derived from the role description, meaning they focus on what is explicitly required avoid mentioning irrelevant points from the file content that are not present in the role description avoid making mistakes avoid making things up, it is very dangerous to ask a question that contradict the file content use a schema as bulletpoints: - question - purpose of the question - mini quote file content - mini quote role description --- refine questions, ensuring they comply with my instructions focus on the requirements of the role description validate consistency of the questions with the provided information output only the list of refined questions avoid any comments at the beginning or the end of your message --- now, i will present you with a number followed by some text number indicates the question you just created text is a quick note of the answer of the person that i am taking during the conversation routine - repeat your question - rewrite the answer in a more coherent way without adding new meaning and stay close to the answer - assess the answer on the terms of - factual correctness - search the web if necessary - completeness - did the applicant miss something important - relevance - does it correspond to what was being asked - application - if relevant, does the answer consider real-world implications or applications - evidence: did the applicant cite relevant sources or provide examples to support their points - give your verdict about the assessment of the answer - decide pass / no pass - short explanation - propose a follow-up question - to dig deeper on the same topic, or - to explore tangential topics - alternatively, change topics completely - if faced with a new question as indicated either by a number at the beginning or a double question mark like this ??, you should restart the cycle like this - forget about the previous question - repeat the routine - special clause: if some later information is relevant to your assessment, remind me about it next - analyze these instructions - write a very brief summary of what you are going to do - send a message stating you understood - wait for my input ---