PhD Defense Mock Q&A Dossier

Template Overview

Defending a doctoral dissertation before an academic committee is an intellectual trial that demands poise, deep domain fluency, and rigorous defense of methodological decisions. The PhD Defense Mock Q&A Dossier allows doctoral candidates and their advisors to catalog expected challenges from committee members, map thesis chapters to theoretical foundations, rehearse defensive responses, and shore up potential intellectual vulnerabilities.

Markdown Template Structure

Copy the block below or click Download .md File to add this template directly to your personal vault:

---
title: "Defense Dossier: [Dissertation Title]"
candidate_name: "[[PhD Candidate Name]]"
defense_date: [YYYY-MM-DD]
dissertation_advisor: "[[Prof. Advisor Name]]"
external_examiner: "[[Prof. External Name, Institution]]"
committee_members: ["[[Member 1]]", "[[Member 2]]", "[[Member 3]]"]
tags: [phd-defense, dissertation, doctoral, academic, viva-voce, defense-prep]
---

# PhD Defense Mock Q&A Dossier — [Dissertation Title]

## 1. The 3-Minute Thesis Defense Summary
- **The Core Problem:** Digital knowledge workflows face a tension between AI assistance and intellectual sovereignty.
- **The Thesis Claim:** High-efficiency, quantized on-device neural language models provide computational assistance equal to cloud endpoints while eliminating privacy vulnerabilities.
- **Primary Novel Contribution:** A mathematically proven attention-pruning heuristic specifically optimized for unified memory edge chips.

## 2. Committee Persona & Intellectual Bias Profile
| Committee Member | Academic Specialization | Typical Questioning Philosophy | Identified Point of Vulnerability |
|---|---|---|---|
| Prof. Advisor | Cognitive Systems | Friendly, supportive, probes future work | Wants to see clear postdoctoral roadmap |
| Prof. External Examiner | Mathematical Foundations | Skeptical, demands formal proofs and bounds | Will challenge the mathematical convergence rate in Chapter 4 |
| Prof. Systems Lead | Operating Systems & Compilers | Pragmatic, focuses on hardware benchmarks | Will question memory overhead during real-world thermal throttling |

## 3. High-Risk Questions & Rehearsed Responses
### Question 1: Theoretical Generalizability (Chapter 3)
- **The Challenge:** "Your evaluation relies on Apple Silicon architecture. How do your mathematical conclusions generalize to x86 or RISC-V edge cores?"
- **Rehearsed Defense Strategy:**
  1. Acknowledge architecture boundaries clearly: the empirical latency figures reflect unified memory throughput.
  2. Pivoting to theoretical foundation: The mathematical attention pruning proof (Lemma 4.2) depends on token sparsity, which is hardware-agnostic.
  3. Citing benchmark cross-validation: Point to preliminary benchmarks on Snapdragon chips demonstrating identical algorithmic reduction.

### Question 2: Scope Limitation (Chapter 5)
- **The Challenge:** "Why did you exclude real-time multi-speaker diarization from your primary system scope?"
- **Rehearsed Defense Strategy:**
  1. State that clustering audio embeddings on-device without cloud anchors introduces quadratic memory expansion that violates the single-chip latency ceiling.
  2. Frame as deliberate thesis scope boundary: the core research question centered on transcription accuracy and zero network transmission.

## 4. Key Citations & Mathematical Armor
- **Lemma 4.1 Citation:** Finch & Rostova (2025), *Journal of Machine Learning Research*.
- **Baseline Comparison:** Touvron et al. (2023) LLaMA attention-head caching metrics.

## 5. Defense Day Protocol & Countdown Checklist
- [ ] Final draft printed and bound with chapter tab markers.
- [ ] Offline slide presentation tested on presentation room projector.
- [ ] Water bottle, timer, and physical legal pad for jotting committee questions.

Filled Real-World Example

Here is a realistic example demonstrating how this template captures actual deliberations:

---
title: "Defense Dossier: Edge Transformers on Unified Memory"
candidate_name: "[[Dr. Samantha Wei]]"
defense_date: 2026-10-09
dissertation_advisor: "[[Prof. Julian Vance]]"
external_examiner: "[[Prof. David Thorne, Oxford]]"
committee_members: ["[[Prof. Elena Rostova]]", "[[Prof. Marcus Chen]]"]
tags: [phd-defense, dissertation, doctoral, academic, viva-voce, defense-prep]
---

# PhD Defense Mock Q&A Dossier — Edge Transformers on Unified Memory

## 1. The 3-Minute Thesis Defense Summary
Centralized speech AI forces a false dilemma between modern productivity and confidentiality. My dissertation proves that 4-bit edge transformers running on consumer unified memory achieve parity with cloud APIs, establishing the theoretical and empirical foundation for zero-leakage local AI tools.

## 2. High-Risk Questions & Responses
- **Prof. Thorne (External Examiner):** "Why should the research community care about quantization when consumer hardware memory grows larger every year?"
- **Defense Response:** "Memory capacity may increase, but memory bandwidth and thermal ceilings remain strict physical bottlenecks on battery-powered edge hardware. Pruning attention caches is essential for sustained real-time performance regardless of total RAM size."

## 3. Defense Day Checklist
- [ ] 20-minute slide deck timed to exactly 18:30 during final dry run.
- [ ] Backup USB flash drive containing standalone PDF slide export.
- [ ] Confirm room key access with department administrator.

When to Use and When to Avoid

When to Use

• Use during the final 3 months leading up to a doctoral dissertation defense (viva voce).
• Use during mock defense rehearsals with lab peers and academic advisors.
• Use to catalog and stress-test candidate responses against known committee biases.

When to Avoid

• Avoid using during the early proposal phase of a PhD (use doctoral supervision log).
• Avoid using for brief undergraduate classroom presentations.

On-Device AI Integration

Summarize your recording in Oyma with this local instruction:

Analyze the PhD defense mock dossier:
1. Identify weak points in the candidate's planned responses where committee members could press further.
2. Formulate two hostile follow-up questions challenging the novelty of the primary contribution.
3. Suggest rhetorical reframing strategies to transform scope limitations into rigorous methodological discipline.
Output in structured Markdown.

Your data remains securely on your Mac with zero cloud exposure.