Skip to content
tezvyn:

🤖AI & ML

Artificial intelligence, machine learning, and data science

308 bites

Test yourself: Top 30 advanced AI & ML interview questionsMultiple choice, with the correct answer and why it is correct on every question. Free, no sign-in.

Advanced everything in AI & ML

advanced1 min read

Automated rollback for a failed model deploy

Detect failure via health and metric checks, automatically route traffic back to the last known-good version (blue-green or canary), keep the registry entry but unpromote, and alert.

advanced1 min read

CI/CD for microservice-based ML systems

Independent per-service pipelines, contract testing to protect interfaces and schemas, and incremental deploys (canary, blue-green); manage data and model contracts, not just code.

advanced1 min read

Fairness and robustness gates in CI/CD

Sliced fairness metrics across subgroups, robustness checks via perturbation and adversarial sets, all compared to thresholds that fail the build.

advanced1 min read

Managing model-as-a-feature pipelines

An upstream embedding model becomes a versioned dependency, creating cascading retraining, version skew, latency stacking, and lineage complexity.

advanced1 min read

Scalable multi-modal data quality pipeline

Staged distributed pipeline doing schema and integrity checks, modality-specific filtering, dedup, PII and toxicity removal, and metric-gated quarantine.

advanced1 min read

Three techniques to cut LLM inference latency

Quantization shrinks weights with small accuracy risk, KV-cache plus continuous batching boost throughput, speculative decoding drafts tokens for lossless speedup.

advanced1 min read

Differential privacy vs utility in LLM fine-tuning

Clipping plus calibrated noise per step, smaller epsilon means stronger privacy but degraded accuracy, tracking the privacy budget across epochs.

advanced1 min read

Core insight behind GPTQ and AWQ

Not all weights matter equally; GPTQ minimizes layer output error using second-order info, AWQ protects salient weight channels tied to large activations.

advanced2 min read

Tensor versus pipeline parallelism for large models

Tensor parallelism splits individual layers across GPUs needing fast interconnect; pipeline parallelism splits layers into stages across GPUs.

advanced1 min read

Setting up an LLM-as-a-judge evaluation

Define rubric, prefer pairwise comparison, randomize order, calibrate against humans; control position, verbosity, and self-preference bias.

advanced1 min read

Batching strategy for multimodal training

Control dataset mixing ratios, use balanced sampling and per-source weighting, keep enough text-only data to avoid forgetting, and handle variable shapes via grouping or padding.

advanced1 min read

LLaVA versus Flamingo vision-LLM design

LLaVA projects image features into the LLM input space and feeds them as tokens, keeping the LLM mostly intact; Flamingo inserts gated cross-attention layers inside a frozen LLM.

advanced1 min read

Temporal consistency in video diffusion

Add temporal layers, such as temporal attention or 3D convolutions across frames, so the model attends across time and frames denoise jointly rather than independently.

advanced1 min read

Diffusion-based image inpainting design

At each denoising step keep the known region by replacing it with the noised original, let the model generate only the masked area, condition on prompt and mask.

advanced1 min read

DDIM: faster diffusion sampling

DDIM defines a non-Markovian deterministic process sharing DDPM's training, letting you skip steps and sample in far fewer iterations.

advanced1 min read

Designing an agent that resolves ambiguity

Detect ambiguity, gather evidence with the contact API, resolve relative time deterministically, ask the user only when genuinely uncertain, then confirm before the irreversible booking.

advanced1 min read

Securing tool-using LLM agents

Name indirect prompt injection, data exfiltration, and unsafe tool execution, then defend with sandboxing, least-privilege scoped tools, input/output filtering, and human-in-the-loop on risky actions.

advanced1 min read

Hybrid search and re-ranking for retrieval

Hybrid search fuses dense semantic and sparse keyword signals to catch exact terms dense misses; a cross-encoder re-ranker rescoring top-k boosts precision.

advanced1 min read

Direct Preference Optimization explained

DPO reparameterizes the RLHF reward in terms of the policy itself, turning alignment into a simple classification loss on preference pairs with no separate reward model or PPO.

advanced2 min read

Federated learning architecture, risks, and defenses

Devices train locally and send updates not data, a server aggregates; gradients still leak, enabling inversion or membership inference; defend with secure aggregation and DP.

We are hiring for this. Every open role lists the topics its interview covers, so you can prepare for the real thing rather than guessing.

See open roles