Co-founder, Acyclic Labs (YC F26). Rapid agent swarms: make any agentic workflow exponentially faster. ex-Palantir, Cambridge CS.
london / sf
Joined May 2022
- Tweets378
- Following265
- Followers626
- Likes572
Pinned Tweet
im 23, in sf for tech week, building agent swarms at acyclic (yc f26)
if you're here and building something real, let's connect
Oooo this one looks pretty fire
Releasing the model weights of Kimi K3.1 Pro: huggingface.co/moonshotai/Ki…
We'll keep advancing open intelligence.
check this out, their mission is to solve all of engineering. some real star trek ts
Today we’re launching @vorelios (YC F26)!
We’re a research lab building AI that can engineer anything. Models that predict a design’s behavior in seconds and use those predictions to invent and improve designs autonomously. 🧵
w/ @krish_k07 , @ycombinator
vorelios.com
Ram Vinjamuri retweeted
we're releasing a free plugin to render MIDI into continuous audio.
get natural sounding strings, brass, wind and guitars via just MIDI inputs to our plugin that lives in the DAW.
try out the beta today at midi.audio
Ram Vinjamuri retweeted
so close to 500 followers
if you're building with agents, in sf, or want to see me try to ship agent swarms in public, hit follow
i'll follow you back if you're not a bot
followers friday 🫡
went from ~430 to 606 followers this month, mostly from the demo videos. last few posts were mid ngl
building agent swarms at acyclic.dev in public, wins and Ls
if you're into agents, infra or evals, drop your handle below and i'll follow back (if you're not a bot)
flicked through one of my prs and found this
can someone explain how google's api keys actually work? GOOGLE_API_KEY vs GEMINI_API_KEY
and i can run gemini and gemma on gcp and on google ai studio. what's the difference, and why aren't they the same? (or are they?)
Ram Vinjamuri retweeted
This is our YC launch.
I talk about armies of tiny servers.
you should watch it.
we are at the frontier but you can't use it cause we are pacing the frontier
Carbon is rolling out first to cyber defenders and we are working to make it more widely available soon!
blog.google/innovation-and-a…
here at @kaggle @GoogleDeepMind bench summit today
we're building benchmarks for swarm coordination, and for the pitfalls you hit when thousands of agents run at once
say hi if you're around
mocking tests is a problem which is getting bigger and bigger as products integrate with more and more apps and these guys are building at the frontier of the space, making it easier for everyone
Three months ago @ev_whitehouse and I set out to fix how agents get tested. Mocks aren't real enough and sandboxes don't scale.
Today we're launching Rystic (YC F26) : local copies of Slack, Stripe, S3 and the rest of your stack.
→ Save scenarios
→ Inject faults
→ Run at scale
Check it out below 👇
after the blacksmith raise, genuinely curious: what are people actually running CI on?
mine: blacksmith, 7/10. fast but expensive, even with the yc discount
it's 11pm on a friday in sf
everyone i know is out. i'm here with a whiteboard figuring out how to make 200 agents migrate a django app even faster than 49 minutes for $8
quit my cushy 6 figure job for this
late late night. who else is still building?
Demo #5: we pointed 200+ agents at a real e-commerce codebase and migrated it from Django 1.11 to 4.2.
Root → 15 orchestrators → one agent per file; 100s at once.
49 minutes. ~$8
Every transcript is public 👇
acyclic.dev/evals/examples/o…
Ram Vinjamuri retweeted
Shipping code shouldn’t mean hours of manual testing.
@Simulithic (YC F26) sends AI agents through your app like real users, across browsers, devices, and account types. It tells you what broke without you writing a single test script.
The teams that ship fastest will be the ones that can verify as fast as they build.
Get started today: simulithic.com
feels like everyone's token bills have crept up lately, curious where people are at
i'll go first: ~$3k/month, mostly on API and coding subscriptions. heavily subsidised by having way too many subs, one of my claude code plans burned the equivalent of $2.5k in API
reply with your spend (personal or company, ranges are fine) and what's eating it. coding agents, long-running agents, doc pipelines, something else?