The design library
Building blocks, complete architectures, and variations worth comparing. Search by the problem you want to solve or a component you want to understand.
Classic designOfficial
Short clips from a prompt: latent video diffusion, recaptioned data, spacetime attention, cascades, multi-GPU jobs, previews, fair queues, provenance: one-hour boards for junior, senior and staff, with the theory behind them.
Classic designOfficial
Images at 2K–4K without paying for every pixel: latent diffusion, an SR cascade with noise augmentation, tiled decoding, distilled drafts, cost per megapixel: one-hour boards for junior, senior and staff, with the theory.
Classic designOfficial
Selfies in, professional headshots out: per-order LoRA on SDXL, an identity-encoder preview, face-match ranking, GPU pools, consent and deletion: one-hour boards for junior, senior and staff, with the theory behind them.
Classic designOfficial
Faces of people who do not exist: a style-based GAN, truncation, latent-space edits, selfie inversion, provenance and memorisation audits: one-hour boards for junior, senior and staff, with the theory behind them.
Classic designOfficial
Captions read aloud as alt text: encoder, bridge and decoder, CLIP-filtered data, beam search, hallucination-aware evaluation, lanes on one GPU pool: one-hour boards for junior, senior and staff, with the theory behind them.
Classic designOfficial
Prompt to four images in seconds: latent diffusion, guidance, a distilled few-step student, recaptioned data, safety on both sides, watermarks and C2PA: one-hour boards for junior, senior and staff, with the theory behind them.
Classic designOfficial
Translation for 20 to 240 languages: one multilingual encoder-decoder, mined and back-translated data, distilled students, COMET and MQM, quality-aware serving: one-hour boards for junior, senior and staff, with the theory.
Classic designOfficial
Next words within 100 ms of a keystroke: conditional language modelling, a distilled student on draft-pinned KV caches, private n-grams, DP and canary audits: one-hour boards for junior, senior and staff, with the theory behind them.
Classic designOfficial
Answers from private documents with citations: chunking, BM25 + vector retrieval, reranking, permission filters, grounding checks and evaluation: one-hour boards for junior, senior and staff, with the theory behind them.
Classic designOfficial
Teacher models write training data for smaller students: conditioned and evolved prompts, verified answers, dedup, decontamination, versioned datasets with lineage: one-hour boards for junior, senior and staff, with the theory behind them.
Classic designOfficial
Every new LLM version shrunk to FP8, INT8 or INT4 with a smaller KV cache, gated against its BF16 parent and benchmarked on the serving GPU: one-hour boards for junior, senior and staff, with the theory behind them.
Classic designOfficial
A 70B teacher distilled into an 8B student: transfer sets, logit and on-policy distillation, per-slice gates, escalation routing, a refresh loop: one-hour boards for junior, senior and staff, with the theory behind them.
Classic designOfficial
Tenants upload examples and get a tuned, gated model on the same API: LoRA and QLoRA arithmetic, chat templates, Kueue fair sharing, eval gates, multi-LoRA serving: one-hour boards for junior, senior and staff, with the theory behind them.
Classic designOfficial
Preference tuning as a weekly loop: rater and AI labels, Bradley–Terry reward models, DPO and PPO with a KL leash, vLLM rollouts, reward-hacking checks: one-hour boards for junior, senior and staff, with the theory behind them.
Classic designOfficial
Pretraining a base LLM on a fixed budget: sizing by 6ND and scaling laws, the data factory, tokenizer, stable runs, failures and evals: one-hour boards for junior, senior and staff, with the theory behind them.
Classic designOfficial
A chat assistant on models we train and serve: next-token framing, three training stages, resumable SSE streams, prefix caching, token quotas, safety, evaluation: one-hour boards for junior, senior and staff, with the theory behind them.