Aashi Agrawal

swe @ DoorDash | prev @ Notion, Apple | uiuc cs alum

[Aashi Agrawal @aashi_wal]

if you already spend half your day texting, why not order food that way too? I built MunchMate - an AI agent that takes your order via text and places it on DoorDash for you🤖🍔

huge shoutout to @browserbase + @pk_iv - their browser automation stack is insanely powerful ⚡️

[Aashi Agrawal @aashi_wal]

@adelwu_ My brown reducto shirt is the best company swag I own fr

[Aashi Agrawal @aashi_wal]

@hahnbeelee Built a chatbot last year to order off DoorDash😋

[Aashi Agrawal @aashi_wal]

probably the first late night cafe event where I actually got work done + met the coolest engineers (plus got insane merch)

thanks for hosting @adelwu_ 🤩

[Aashi Agrawal @aashi_wal]

@adelwu_ @reductoai So hyped🤩

[Aashi Agrawal @aashi_wal]

@adelwu_ Would love to attend and catch up with you again!!

[Aashi Agrawal @aashi_wal]

Software sprawl became agent sprawl

[Perplexity @perplexity_ai]

Introducing Perplexity Computer.

Computer unifies every current AI capability into one system.

It can research, design, code, deploy, and manage any project end-to-end.

[Aashi Agrawal @aashi_wal]

@saharsh Cool article! p much RL but for startups

[Aashi Agrawal @aashi_wal]

luma-maxing is all fun and games until you start seeing the same people over and over again and realize you’re lowk in a cult

still fun tho

[Aashi Agrawal @aashi_wal]

@derekmeegan I was literally thinking this today

[Aashi Agrawal @aashi_wal]

https://t.co/LD0ImCG1AX

[Aashi Agrawal @aashi_wal]

Read this paper on DISK feature extractors and it shifted my view on why learning local features is hard. A lot of prior work optimized the descriptor itself, but the real challenge is deciding where to look and what to match. DISK uses RL to learn these decisions

[Annie ❤️‍🔥 @AnnieLiao_2000]

Enterprise AI adoption will be the hottest topic of 2026.

We pre-empted this, and are launching Solaris: a platform to help enterprises understand what’s possible with AI and actually make it happen.

We are already live with fast growing companies like Lendi Group, Eucalyptus.

And are partnered with incredible companies like @ManusAI, @nebiusai, @juliusai, @getlindy etc. In 2026, we are partnering with @HP in APAC to power the future of the AI workforce.

More below

[Aashi Agrawal @aashi_wal]

Used to think image matching was a solved problem - just match points btwn two images.
Turns out it’s still one of the hardest problems in CV 😵‍💫
This paper explains why even modern methods like LoFTR still struggle and where research is headed next
https://t.co/vGy5wEa6uE

[Aashi Agrawal @aashi_wal]

@github pls fix urself I need to deploy my code otw I get fired😭

[Aashi Agrawal @aashi_wal]

Love a self-aware model

[Aashi Agrawal @aashi_wal]

PR inflation at work goes crazy
I was talking to an engineer who said they ship faulty code and then ship a fix in a separate PR just to meet their PR count for the week😭

[Aashi Agrawal @aashi_wal]

@nizzyabi @stripe This is just Brex logo but filled in

[Saharsh Agrawal @saharsh]

Today we’re introducing the world's first AI Workforce for Marketing Ops: a team of AI agents that converts 30% more inbound into demos/ signups.

Marketing teams spend millions on inbound, but 50%+ of leads never convert b/c they're lost to long forms, poor targeting, and slow follow-ups.

Here’s how the AI Workforce works →
🧵

[Aashi Agrawal @aashi_wal]

Surface agents 🔛🔝

[Y Combinator @ycombinator]

.@Surfacelabs_ai is building the AI workforce for Marketing Ops.

Surface converts 30% more leads with agents that prevent ad fraud, optimize forms, and route leads to sales. Their agents now manage 50k+ leads per month.

https://t.co/OnHjl0qs5l

Congrats on the launch @saharsh & @arjsahai!

[Aashi Agrawal @aashi_wal]

the way I’m more productive working on code at the airport than I am at home… guess I gotta go on more trips