tarun

eng @turbopuffer | ex: @dragonflydbio , @gitpod , @BuoyantIO

[tarun @tarrooon]

pretty good weather to be in nyc rn!

[tarun @tarrooon]

@davidcrawshaw Oh, this will be so good! Currently using email from exe for notifications!

I'm on Android btw

[tarun @tarrooon]

@turbopuffer puf

[Simon Eskildsen @Sirupsen]

turbopuffer crossed $100M run-rate in March. 19mo after $1M. Profitable & <$1M raised.

Cursor・Anthropic・Notion・Cognition・Harvey・Bridgewater・Ramp・Linear・Legora・Superhuman・Atlassian・Granola

We’d be nowhere without them. We work like hell to exceed their expectations.

[turbopuffer @turbopuffer]

SID-1 is an agentic search model by @SID_AI

→ 1.9x recall over RAG + rerank
→ 24x faster, 99% cheaper than GPT-5.1

trained using large-scale RL on turbopuffer at 1k+ QPS bursts over 10M+ document corpora across thousands of steps

https://t.co/hqmdPUmdLt

[Sualeh Asif @sualehasif996]

A conversation with @sirupsen on scaling Shopify, building turbopuffer, and the future of databases.

0:00 - Scaling Shopify through flash sales and outages
8:13 - How top infrastructure teams collaborated in the 2010s
10:35 - Engineering principles from Logrus and on-call
17:38 - The story behind Simon’s famous-ish blog, Napkin Math
23:05 - Why new database companies keep winning
32:21 - How Simon became a fan of databases
35:45 - AI coding, and where agents still fail
42:10 - Hiring P99 engineers in the AI era
48:45 - What’s next for databases

[tarun @tarrooon]

someone said @turbopuffer is the fastest infrastructure company buildout they've seen, and we're not kidding!

[turbopuffer @turbopuffer]

new tpuf regions, so you can puff a little closer to home

🇧🇷 aws sa-east-1 (São Paulo)
🇺🇸 gcp us-east1 (South Carolina)
🇧🇪 gcp europe-west1 (Belgium)

[Tomás Senart @tsenart]

Starting to map the graph for a founding engineer at Perfloop.

Looking for real systems depth - perf, infra, observability, AI infra, devtools, databases, runtimes - plus extreme agency, product/architecture judgment, and native agentic building.

Europe / near-Europe overlap preferred.

[tarun @tarrooon]

THIS TEAM PUFF'S

[turbopuffer @turbopuffer]

⛷️ turbopuffer summit #004 🏂

our remote team spent a whole week together in banff national park. we brought a camcorder because life is short and why not:

[turbopuffer @turbopuffer]

we couldn’t afford the sphere so we got the next best thing

costs less, hits harder

[nathan 雷迅 @nxlouie]

1/ When we launched Notion AI Q&A in Nov 2023, we knew we were embarking on an ambitious journey.

What we didn’t anticipate was how quickly we’d need to scale—or how much we could optimize costs.

Notion vector search over the past 2 years: 10x scale, then 1/10th cost. 🧵

[turbopuffer @turbopuffer]

we’ve reduced query prices by up to 94%, thanks to infrastructure improvements we've made under the hood

puff harder, for much less

[Will Kenney @willkenney]

what if we dove into search infrastructure together under the fish?

[tarun @tarrooon]

there's so much alpha in co-designing the db and infra together

[tarun @tarrooon]

some amount of `fuck it` energy is necessary in life

[tarun @tarrooon]

turbopuffer is looking for a security engineer or an infra person with a special interest in security! If you know someone who could be a good fit, would appreciate you sharing the link below or DM me! 🤝

https://t.co/3LvZJ4aFnU

[turbopuffer @turbopuffer]

tpuf ANN v3, for when you need to index the entire web

100B+ vectors @ 50ms p50 / 200ms p99 latency

blog: https://t.co/p8naYfgpWe

[turbopuffer @turbopuffer]

december changelog

https://t.co/dYfZZ2mzg9

[tarun @tarrooon]

time moves different at turbopuffer! excited for the next year!

[turbopuffer @turbopuffer]

turbopuffer 2025:
>10x revenue
~5x headcount
+trillions of documents/vectors
+Customers: Anthropic, Linear, Atlassian, Superhuman, Grammarly
+Investors: Thrive, Lachy Groom

[turbopuffer @turbopuffer]

yesterday: short human queries, scalar CPUs
today: long LLM queries, wide SIMD lanes

for FTS v2, we use a vectorized MAXSCORE algorithm instead of WAND, because dumb & serial beat smart & random algorithms on modern CPUs

https://t.co/pXVGycqX7v