What We Are Building
Simulate a real Oxbridge-style interview.
In reality:
•
Students are asked unfamiliar, concept-driven questions
•
Interviewers probe deeper through follow-up questions
•
Students are expected to think aloud and justify their reasoning
ibGuru recreates this process digitally.
The Problem We Are Solving
Most students:
•
Practise interview questions passively
•
Memorise answers instead of thinking
•
Have no access to structured follow-up questioning
•
Do not receive meaningful feedback on their reasoning
As a result, students:
•
Struggle when questions are extended or redirected
•
Cannot defend their answers under pressure
•
Lack depth in explanation
Our Approach
We treat interview preparation as:
A thinking process under pressure, not a set of model answers
Each question should simulate:
Question → Follow-up → Deeper probing → Explanation
Students must:
•
Think through problems step-by-step
•
Explain their reasoning clearly
•
Respond to follow-up challenges
Past Interview Questions (Core)
Each entry should include:
•
Initial question
•
Follow-up questions (linked and progressive)
Follow-ups should:
•
Build on the student’s previous answer
•
Probe assumptions
•
Increase difficulty or abstraction
This creates a chain of reasoning, not isolated questions.
Organisation
Questions should be organised using:
•
Tags (concept-based)
◦
e.g. Membrane Transport, Thermodynamics
•
Source (college-specific)
◦
e.g. Oxford – Oriel – 2015
This allows:
•
Concept-based practice
•
Realistic interview exposure
Model Answers (Supporting Layer)
We will provide model answers for each question.
These should:
•
Demonstrate clear reasoning steps
•
Show how to approach the problem
•
Avoid being overly polished or memorisable
AI can generate initial drafts, but must be:
•
Reviewed
•
Refined by mentors
AI Feedback System
ibGuru replaces the interviewer by providing feedback.
After each response, the system should:
•
Evaluate clarity of reasoning
•
Identify gaps in understanding
•
Suggest improvements
More importantly, it should:
•
Ask follow-up questions
•
Challenge incomplete answers
Video Solutions (Future Layer)
We do not require video solutions for all questions.
However, over time:
•
Mentors may contribute walkthrough videos
•
These should demonstrate real-time thinking
These videos act as:
“How a strong student thinks through the problem”
Notes (Subject-Specific Guide)
Notes should act as a pre-interview lesson, not a blog.
They should:
•
Explain how interviews work for that subject
•
Break down common question types
•
Teach how to approach unfamiliar problems
•
Show how to structure thinking aloud
These should be:
•
Subject-specific
•
Structured like a lesson
•
Consistent across students
Lectures
Lectures may include:
•
Mentor walkthrough of key question types
•
Explanation of interview strategy
•
Discussion of common pitfalls
These should align closely with the Notes.
Role of Mentors
•
Refine AI-generated model answers
•
Improve follow-up questions
•
Ensure questions reflect real interview standards
•
Contribute walkthrough videos when possible
AI is used for generation and scaling,
but mentor input ensures authenticity.