🤖 AI &amp; ML

How FID is calculated versus Inception Score

FID fits Gaussians to Inception features of real and fake images then measures Frechet distance; it uses real references and detects mode collapse.

Cross-attention for visual question answering

In cross-attention queries come from one modality and keys/values from the other, e.g. text queries attend over image features so the question selects relevant regions.

Design a multi-model serving architecture for infrequently used models

Tests sparse-traffic cost efficiency via shared compute and dynamic loading. Strong answers: LRU cache on fast storage, scale-to-zero with async preload, pinned hot models, per-model quotas.

What is cross-validation and why is it more robust than a holdout split?

A single split is noisy and wastes data; k-fold rotates each fold as test, averages scores, and trains on all data.

How to evaluate a RAG system end to end

Measure retrieval (recall, precision, MRR, NDCG) and generation (faithfulness, answer relevance) separately, plus end-to-end correctness.

Why ViTs need positional embeddings

Self-attention is permutation invariant so patch order is lost; positional embeddings restore spatial location. CNNs encode position implicitly via the fixed convolution grid.

easy2 min read

Differences between monitoring a traditional REST API and a production ML model

Contrast latency/errors with ML signals like data drift and training-serving skew against baselines, noting ground truth delays.

High ROC-AUC but low PR-AUC: what does this imply?

Tests if ROC-AUC hides imbalance while PR-AUC exposes it. Severe imbalance dilutes FPR across many negatives, inflating ROC-AUC, but precision crashes. Critical for rare positives with costly false positives. Praising the model on ROC-AUC alone fails.

easy1 min read

Detecting data drift on a continuous feature

Data drift is when serving feature distributions shift from training; detect with a Kolmogorov-Smirnov test comparing distributions; a small p-value signals drift to alert on.

How MMLU works and the contamination problem

MMLU is multiple-choice across 57 subjects scored by accuracy; contamination means test items leaked into pretraining, inflating scores.

Core principles of a Neural Radiance Field

An MLP maps a 3D point plus view direction to color and density; novel views render by casting rays, sampling points, querying the MLP, and volume-integrating along each ray.

K-Means vs DBSCAN: which for geospatial hotspots?

Tests matching algorithmic assumptions to data structure. K-Means needs K and assumes spheres; DBSCAN discovers arbitrary density shapes and labels noise. Choose DBSCAN for geospatial hotspots because density is irregular.

Reference-free evaluation for open-ended dialogue

ROUGE punishes valid paraphrases; use reference-free LLM-as-judge or learned scorers rating coherence, relevance, and groundedness.

advanced1 min read

Interpreting a black-box gradient boosting model

Global tools like permutation importance or aggregated SHAP rank overall feature influence; local tools like per-instance SHAP or LIME explain one prediction; SHAP unifies both via additive…

Pure ViT vs hybrid CNN-Transformer for medical segmentation

Pure ViT captures global context but is data hungry and weak on local detail; hybrid CNN-Transformer gets local features cheaply plus global attention, ideal for scarce…

Design a system to monitor a real-time prediction service for feature drift

Async feature logging, distribution comparison via PSI/KS against training baseline, and threshold-based anomaly alerts.

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.

Attention in diffusion U-Nets for text conditioning

Self-attention mixes spatial features at low-res blocks; cross-attention has image queries attend to text-token keys/values; placed inside transformer blocks.