CTO Emerging Technologies & Innovation 2 — Questions and Answers
Question 1: A CTO wants to evaluate quantum computing's near-term applicability. Which problem class is MOST likely to benefit from near-term noisy intermediate-scale quantum (NISQ) devices?
- Large-scale relational database queries
- Combinatorial optimization problems like portfolio balancing (Correct answer)
- Real-time video transcoding
- Distributed ledger consensus mechanisms
Correct answer: Combinatorial optimization problems like portfolio balancing
NISQ devices are best suited for combinatorial optimization tasks where even approximate quantum speedup offers business value.
Question 2: Which edge computing deployment model places compute resources at the network's outermost boundary—closest to sensors and end devices—rather than at regional aggregation points?
- Fog computing
- Cloudlet computing
- Far-edge (device-edge) computing (Correct answer)
- Multi-access edge computing (MEC)
Correct answer: Far-edge (device-edge) computing
Far-edge or device-edge computing places processing directly on or immediately adjacent to IoT sensors and end devices.
Question 3: A startup proposes using a large language model (LLM) to autonomously execute production database changes. As CTO, what is the PRIMARY governance risk to address?
- Model inference latency exceeding SLA thresholds
- Lack of deterministic rollback and human-in-the-loop approval gates (Correct answer)
- Insufficient GPU memory for large schema introspection
- Token context window limitations for long SQL scripts
Correct answer: Lack of deterministic rollback and human-in-the-loop approval gates
Autonomous LLM-driven production changes require deterministic rollback capabilities and mandatory human approval to prevent irreversible data loss.
Question 4: In a digital twin implementation for a manufacturing plant, which data characteristic is MOST critical for the twin to provide real-time anomaly detection?
- Historical batch exports updated nightly
- High-fidelity sensor streaming with sub-second latency (Correct answer)
- Aggregated weekly KPI dashboards
- Manual operator input logs
Correct answer: High-fidelity sensor streaming with sub-second latency
Real-time anomaly detection in digital twins requires continuous high-fidelity sensor streams with minimal latency to mirror physical system state.
Question 5: Which 5G network slicing characteristic makes it most valuable for a CTO deploying mixed critical and non-critical enterprise workloads on a shared infrastructure?
- It eliminates the need for physical network upgrades
- It provides isolated virtual networks with guaranteed SLA per use-case (Correct answer)
- It doubles spectral efficiency compared to 4G LTE
- It removes dependency on licensed spectrum bands
Correct answer: It provides isolated virtual networks with guaranteed SLA per use-case
5G network slicing creates logically isolated virtual networks, allowing guaranteed latency and bandwidth SLAs for critical workloads while sharing physical infrastructure.
Question 6: A CTO is assessing serverless computing for a batch analytics workload running 8-hour jobs. What is the PRIMARY architectural constraint to address?
- Serverless functions lack support for Python runtimes
- Maximum execution time limits make long-running jobs unsuitable without orchestration (Correct answer)
- Serverless platforms cannot access cloud object storage
- Cold start latency makes real-time APIs impractical
Correct answer: Maximum execution time limits make long-running jobs unsuitable without orchestration
Serverless functions have hard execution time limits (typically 15 minutes on AWS Lambda), requiring orchestration patterns like Step Functions for longer workloads.
Question 7: When evaluating a confidential computing solution for sensitive AI model training, which hardware feature provides the strongest isolation guarantee for in-use data?
- TPM 2.0 chip for boot integrity measurement
- Trusted Execution Environment (TEE) such as Intel SGX or AMD SEV (Correct answer)
- AES-256 encryption at rest on NVMe drives
- Hypervisor-level memory ballooning
Correct answer: Trusted Execution Environment (TEE) such as Intel SGX or AMD SEV
TEEs like Intel SGX create hardware-isolated enclaves where even the hypervisor or OS cannot access data being actively processed.
CTO Emerging Technologies & Innovation 2
A CTO wants to evaluate quantum computing's near-term applicability.
Which problem class is MOST likely to benefit from near-term noisy intermediate-scale quantum (NISQ) devices?