AwesomePrep
A candidate rehearses aloud at her laptop, linked by arrows to her CV and job notes, a mock-interview chat, a voice recording and a folder of session files.

A founder story about the practice workflow that led us to AwesomePrep.

Memorised answers fall apart.

An AI mock interview became useful only when it stopped writing answers and started listening, following up and remembering what happened across practice rounds.

Good answers weren’t enough.

A friend received an invitation to interview with a large company. We helped her prepare through three rounds, and she eventually received an offer. This is one person’s experience, not a promise that the same process guarantees an interview result.

We began with sensible preparation: the job description, her CV, the recruiter’s guidance and publicly available information about the interviewers’ work. From that material, we built questions around the role and the kinds of decisions she might need to explain.

Then we drafted answers. They were structured, complete and reassuring to read. They were also much less useful once somebody asked, “Why did you choose that?” or “Which part was actually yours?”

She could remember the prepared story. A different wording or an unexpected follow-up could still break the script.

We changed the practice.

Instead of asking the agent to produce another polished answer, we told it to behave like an interviewer.

  1. Ask one question.She answered aloud, without reading from a prepared script.
  2. Follow the answer.The next question came from what she had just said, including vague claims and missing evidence.
  3. Give immediate feedback.The agent pointed out where the reasoning, ownership or measurement was unclear.
  4. Keep the record.After each round, it organised the questions, answers, follow-ups and feedback into files.

We used a coding agent and a separate recorder to make this work. Voice input handled the live conversation, while the recording preserved what she actually said for later review.

The pattern mattered more than one answer.

A single answer can look like a small wording problem. Several recorded sessions reveal something deeper. We used a stronger model to compare the practice files and look for gaps that kept returning:

  • Did she describe the team’s work without making her own contribution clear?
  • Could she state the result but not explain why she made the decision?
  • Did she name a method without showing how she judged whether it worked?
  • Which follow-up questions repeatedly disrupted her thinking?

Her feedback was that the practice did not help her memorise more answers. It helped her organise her thinking, recognise the key concept in a question and develop an answer from there.

The useful workflow was also a messy one.

Every session required a coding agent and recording software to stay open together. When a round ended, we had to ask the agent to organise the written record and then file the recording ourselves.

That friction became the product idea: put the preparation material, live questions, adaptive follow-ups, recordings, feedback and cross-session analysis in one place.

Practise the conversation, not the script.

AwesomePrep is bringing the complete practice loop into one interface.

Apply for the beta