Step 1 - Three back-to-back online assessments (MCQ-based): • AI/ML Concepts • Analytical & Logical Reasoning • Motivation & Grit Step 2 — Round 1 via FloCareer (third-party platform): The JD loaded on FloCareer was for a Technical Lead / Manager level role, which was completely misaligned — I have 2 years of experience and was told this was open to candidates with 2 YOE as well. The interviewer appeared to be generating questions by pasting the JD into ChatGPT and asking those directly — the questions felt templated and detached from any real evaluation. He also gave a hard-level coding problem, which I was not able to solve. After that round, I contacted the HR directly. He clarified that the JD was "just for reference" and the role is open to early-career candidates too. He was transparent, responsive, and calm throughout — genuinely appreciated that. He arranged a second chance with the internal team. Step 3 — Round 2 with internal interviewer: This round was technically more serious, but had a clear mismatch with the JD. The role was described as Applied GenAI — engineering mindset, ability to use LLMs in production, no requirement to know LLM internals. The interviewer, however, asked deep Python internals (context managers, generators, etc.) and transformer architecture — none of which appeared in the JD. Not a single question was drawn from my resume or past experience. When I asked the interviewer about the team's current projects, he mentioned they are doing POCs using CrewAI because it's simple to get started. That context made the transformer architecture questions feel even more out of place for the actual work being done.