Hiring manager interview
Assessment withdrawn- Persuasiveness
- Weak
- Structure
- Weak
- Role fit
- Assessment withdrawn
- Follow-up handling
- Weak
- Delivery and pacing
- Weak
- Your reverse questions
- Weak
This sample session was conducted in Chinese. Your own interview and report run in the language you choose.
The interviewer asked
Please go into a representative result you delivered in model fine-tuning or quality optimisation, and the metrics used to evaluate it.
What the candidate said (recognized transcript) · Original, verbatim
Persuasiveness“漏洞检测项目,我们标注了六万多条数据,用Lora微调大模型,把静态分析误报率从65%降到25%。另外结合AST和CFG图结构,做代码审计,多任务指令微调加思维链,内部测试准确率到98%。”Translation
On the vulnerability-detection project we labelled more than 60,000 samples and fine-tuned a large model with LoRA, cutting the static-analysis false-positive rate from 65% to 25%. We also combined AST and CFG graph structures for code auditing, with multi-task instruction tuning plus chain-of-thought, and reached 98% accuracy in internal testing.
The problem
The 98% accuracy figure lacks an evaluation set, baseline and failure cases, so its meaning cannot yet be assessed. This does not establish that the experience was fabricated.
How to strengthen this answer
Describe the evaluation-set size, baseline, precision, recall and failure cases so the interviewer can check the result.
Editorial correction: the audio says “ops disaster recovery”; “ops cultivation” is a transcription error, not the candidate’s mistake. The original verdict and role-fit rating are withdrawn, with no replacement score. Original transcript and audio remain available to check. Feedback below has been edited to remove unsupported character judgments.
Disclosure: the answers in this session are text-to-speech test material, not a real person's recording; the role and work history are demo data. The criticisms are model output and have not been independently verified.