CrestingRadarGet featured →

Rank #40 · on radar since 2026-07-03

tessera

From teacher to tiles — a from-scratch LLM distillation & serving engine: custom Triton/CUDA kernels, FSDP distillation, paged-KV continuous batching, speculative decoding, a Rust gateway, a JAX oracle, and interpretability tooling.

Visit homepage ↗cudarustpytorchflash-attention+12GitHub

Momentum

94.9
24h7d

Why it's ranked

Every score decomposes into published factors — the same math for every tool, paid or not. Read the methodology →

Velocity (weighted, cohort-normalized)0.980
Signal decay0.998
Corroboration1.000
Quality gate1.000

Raw signals (30 days)

github · forks5 latest · 2 snapshots
github · stars467 latest · 2 snapshots