CTO Emerging Technologies & Innovation 3 — Questions and Answers
Question 1: Which innovation framework helps a CTO allocate R&D investment across core business improvement, adjacent market expansion, and transformational bets in a structured ratio?
- Jobs-to-be-Done (JTBD) framework
- McKinsey Three Horizons model (Correct answer)
- Gartner Hype Cycle analysis
- Technology Readiness Level (TRL) scale
Correct answer: McKinsey Three Horizons model
The McKinsey Three Horizons model explicitly divides innovation investment into H1 (core), H2 (adjacent), and H3 (transformational) categories.
Question 2: A CTO wants to implement responsible AI practices. Which technique specifically addresses the risk of an ML model perpetuating historical hiring biases in resume screening?
- Differential privacy during training data ingestion
- Fairness-aware machine learning with disparate impact testing (Correct answer)
- Federated learning across distributed HR systems
- Explainability via LIME or SHAP value attribution
Correct answer: Fairness-aware machine learning with disparate impact testing
Fairness-aware ML with disparate impact testing directly measures and mitigates discriminatory outcomes across protected demographic groups.
Question 3: In the context of Web3 technology evaluation, what distinguishes a Layer-2 scaling solution from a sidechain?
- Layer-2 solutions operate without cryptographic proofs while sidechains use ZK proofs
- Layer-2 solutions derive security from the Layer-1 base chain; sidechains have independent security models (Correct answer)
- Layer-2 solutions require Proof-of-Work consensus while sidechains use Proof-of-Stake
- Layer-2 solutions eliminate gas fees entirely while sidechains reduce them partially
Correct answer: Layer-2 solutions derive security from the Layer-1 base chain; sidechains have independent security models
Layer-2 solutions inherit and settle security on the base Layer-1 blockchain, whereas sidechains maintain separate validator sets and independent security assumptions.
Question 4: A CTO is building a technology radar for the organization. In the ThoughtWorks Technology Radar format, what does placing a technology in the 'Assess' ring indicate?
- The technology should be avoided and phased out immediately
- The technology is worth exploring to understand how it will affect your enterprise (Correct answer)
- The technology is recommended for broad adoption across all teams
- The technology is being trialed in low-risk pilot projects only
Correct answer: The technology is worth exploring to understand how it will affect your enterprise
The 'Assess' ring in ThoughtWorks radar indicates a technology worth researching to understand its potential enterprise impact before committing to trials.
Question 5: Which augmented reality deployment consideration is MOST critical when a CTO evaluates AR headsets for industrial maintenance technicians over warehouse workers?
- Display resolution measured in pixels-per-inch
- Hands-free operation, ruggedization, and safety certification for industrial environments (Correct answer)
- Consumer app ecosystem compatibility and social sharing features
- Battery life exceeding 12 hours for overnight shifts
Correct answer: Hands-free operation, ruggedization, and safety certification for industrial environments
Industrial environments require hands-free AR with ruggedized hardware meeting safety certifications (ATEX, IP ratings) that warehouse consumer-grade devices cannot satisfy.
Question 6: A CTO is designing a synthetic data pipeline to augment a scarce medical imaging training dataset. What is the PRIMARY validation requirement before using synthetic data in model training?
- Verifying synthetic images are indistinguishable to human reviewers
- Confirming synthetic data preserves statistical distributions and does not introduce model performance degradation on real test sets (Correct answer)
- Ensuring synthetic data generation costs less than real data collection
- Validating that the generative model uses GAN architecture rather than diffusion models
Correct answer: Confirming synthetic data preserves statistical distributions and does not introduce model performance degradation on real test sets
Synthetic data must preserve statistical fidelity and must be validated against held-out real test sets to confirm it improves rather than degrades model performance.
Question 7: Which characteristic of neuromorphic computing chips makes them particularly advantageous for always-on IoT edge inference compared to traditional GPU-based inference?
- Higher TFLOPS throughput for transformer model inference
- Event-driven sparse computation enabling ultra-low power consumption (Correct answer)
- Compatibility with CUDA-based deep learning frameworks
- Support for 64-bit floating point precision calculations
Correct answer: Event-driven sparse computation enabling ultra-low power consumption
Neuromorphic chips process only when input events occur (event-driven), enabling microwatt-level power consumption ideal for battery-powered always-on edge devices.
CTO Emerging Technologies & Innovation 3
Which innovation framework helps a CTO allocate R&D investment across core business improvement, adjacent market expansion, and transformational bets in a structured ratio?