TickerTain
TickerTain
NewsroomShortsPortfolioConvergence
NewsroomShortsPortfolioConvergence
←
▶ 16:11 · $FACTORY · Sequoia-backed Factory leads security coding benchmarks via Fireworks post-training
episode briefing
Sequoia Capital

When (and How) to Post-Train Your Own AI Models | Lin Qiao, Fireworks

2026-08-12 · 4 company · 10 thematic
sentiment
4 bull0 bear0 neu
speakers
lin qiao

Lin Qiao (Lynn Quo) founded Fireworks AI after 7 years at Meta (Facebook) where she worked on distributed systems and databases. She started the company at age 48 after deliberately building leadership skills. Fireworks provides a specialized intelligence platform for model customization and inference, processing 40T+ tokens/day with $800M ARR.

quote
“Factory, that's another Sequoia company. They build on top of Fireworks and especially focus on security part of the coding. That is a very hard topic because security is not high tolerance. You need to get it right. They tune a model that…”
—lin qiao
now playing · $FACTORY
$CURSORbullish· mediumlin qiao
Cursor achieves frontier-level coding quality via deep post-training on Fireworks
Cursor started post-training early, executes deep mid-to-post training, and its Composer 2.5 model now matches or beats frontier lab quality on coding benchmarks, proving the spec…
$DOCS···bullish· mediumlin qiao
Doximity tops Stanford-Harvard clinical safety benchmark with Fireworks-tuned model
Doximity's clinical AI, trained on Fireworks, achieved top rank on the Stanford-Harvard clinic safety benchmark, demonstrating that domain-specific post-training can surpass gener…
$FACTORYbullish· mediumlin qiao
Sequoia-backed Factory leads security coding benchmarks via Fireworks post-training
Factory, a Sequoia portfolio company, uses Fireworks to post-train a security-specialized coding model that tops security benchmarks, showing that low-tolerance domains like secur…
$JEN-SPARKbullish· mediumlin qiao
Jen Spark beats frontier models on cost with 5-10x reduction via post-training
Jen Spark's generic co-work application, post-trained on Fireworks, slightly outperforms frontier models on quality while achieving 5-10x lower serving cost, proving the unit econ…