← Back
Home › Systems › AI applications

recommender

recommendation system

TikTok, YouTube

SystemTechniques · 16Primitives · 43 recommender Windowed stream aggregationdistributed · medium: Group events by event-time window per key, keep one partial aggregate per slice, and emit a window when the watermark passes its end.›Matrix factorization for collaborat…machine-learning · medium: Learn a short latent vector per row and per column from the observed cells only; a missing cell is predicted by the dot product of its two vectors.›Contrastive learningmachine-learning · medium: Train an encoder so two views of the same thing score higher than every other pair in the batch; InfoNCE with a temperature, no labels needed.›Feature interaction modelsmachine-learning · medium: Click/conversion prediction networks for ranking: feature crosses (DCN, DeepFM), embedding bags, calibrationFeature storemachine-learning · medium: One definition per feature feeds a history store for training and a latest-value store for serving, so training and inference see the same values as of the same moment.›Multi-armed banditsmachine-learning · medium: Choose among K options repeatedly while learning their payoffs: epsilon-greedy, UCB1 and Thompson sampling keep total regret low where a fixed split or a fixed exploration rate pays forever.›Two-tower retrievalmachine-learning · medium: Two separate encoders map users/queries and items into one vector space; item vectors are precomputed and searched by nearest neighbour, so recall over millions costs one small forward pass.›Negative samplingmachine-learning · low: Train by contrasting each observed positive with a few sampled negatives instead of normalising over every candidate; the choice of sampler decides bias and quality.›Embeddingsmachine-learning · low: A learned table mapping discrete ids to dense vectors: gather forward, sparse scatter-add backward, row-wise optimizers to keep huge tables cheap.›Calibrationmachine-learning · low: Make predicted probabilities match observed frequencies: measure with ECE, fix post hoc with temperature, Platt or isotonic scaling on held-out data.›Top-k selectionsearch-sort · low: Keep the k best of n items without sorting all n - bounded min-heap for streams, quickselect for arrays, threshold pruning to skip most items.›Count-min sketchdata-structures · low: d rows of w counters estimate any item's frequency in a stream, overestimating by at most ε·N with probability 1-δ, in fixed tiny memory.›Evaluation metricsmachine-learning · low: Precision, recall, ROC AUC, average precision, log loss, NDCG and pass@k computed correctly, plus paired confidence intervals for comparing models.›Embedding table shardingml-training · low: Sharding huge embedding tables across devices (row/column/table-wise) with all-to-all lookups and cachingTransformer blockml-training · low: The decoder-only pre-norm block (attention plus MLP on a residual stream) behind every modern LLM, with exact shapes, parameter, FLOP and KV-cache counts.›Maximal marginal relevancesearch-sort · low: MMR and other diversification re-ranking of result listsHash aggregationstorage · low: GROUP BY via hash tables with partitioning and spill to diskHierarchical timing wheelsystems · low: Timers bucketed by expiry into wheels of slots, coarser wheels for far deadlines: O(1) insert and cancel for millions of timeouts.›LU and Cholesky factorizationnumerics · low: Factor A = PLU (any square matrix, partial pivoting) or A = LLᵀ (SPD) once in O(n³), then solve each right-hand side in O(n²).›Sparse matrices (CSR)numerics · low: Compressed sparse row stores only nonzeros as row pointers, column indices and values; SpMV streams them once, so it is memory-bandwidth bound.›Data augmentationmachine-learning · low: Generate label-preserving variants of each training example on the fly to enlarge the effective dataset and bake in invariances.›Loss functionsmachine-learning · low: The scalar a model minimizes: cross-entropy for classes and tokens, MSE or Huber for values, each with a stable form and simple gradient.›Feature hashing (the hashing trick)machine-learning · low: Hash each feature name to an index in a fixed-size vector, with a second hash for the sign, so no vocabulary is needed and memory is fixed.›+ 9 more techniques+ 27 more primitives

Similar systems

Share the most building blocks with this one.

realtime-personalization 45 %ads-auction 31 %fraud-detection 25 %short-video 19 %model-monitoring 19 %