Neha Sarda

global youth advisor @Google (via canvas8 - views are my own) | growth engineer @AquinF03

[Neha Sarda @nsardaa]

This chart is actually so satisfying to look at!

deep + cheap quadrant is everything we're going for at @Aquinf03, like why would you spend serious compute on surface level signals.

Aquin isnt bottom right and it shows deep insights with cheap ways to get there!

[Ash @Ashf03]

for the first time you can simulate fine-tuning before committing any compute

with @AquinF03, see exactly which features strengthen or suppress, which samples hurt generalization, which layers go dead!

[Neha Sarda @nsardaa]

@thelaunchd @interact_studio @reactorworld this is so cool WOW

[Neha Sarda @nsardaa]

@Ashf03 @AquinF03 if this isn’t the most goated thing ever!

[Neha Sarda @nsardaa]

@devhunt_ @Ashf03 looks familiar 😝 @AquinF03 @Ashf03

[Neha Sarda @nsardaa]

@thefoundingco_ @theresidency woohoo!!

[Neha Sarda @nsardaa]

@AquinF03 aquin’s the best icl

[Aquin Labs @AquinF03]

If you get in, we'll provide you compute 🦾💪

[Neha Sarda @nsardaa]

@AquinF03 @sammmbhav @MaxMill06 super cool

[Aquin Labs @AquinF03]

If you do mech interp research, @sammmbhav and @MaxMill06 made badges for your repos!

from SAEs, causal tracing, evals to benchmarks and more!

https://t.co/ZMXuZ6XLo4

[Neha Sarda @nsardaa]

@JaswanthBeere @AquinF03 @VivGadepalli @Ashf03 @sammmbhav @pcannon775 @MaxMill06 @deascrap how sweet!

[Neha Sarda @nsardaa]

@AquinF03 hands down the coolest

[Aquin Labs @AquinF03]

We used Aquin to debug and improve a QLoRA fine-tuning pipeline for LLaMA 3.2 1B on a healthcare dataset.

With a sdk to transform any local into communicable server for the application!

[Neha Sarda @nsardaa]

@Ashf03 @AquinF03 @sammmbhav @pcannon775 @MaxMill06 everyday thing :(

[Neha Sarda @nsardaa]

@deascrap @AquinF03 lesgooo

[Neha Sarda @nsardaa]

@Ashf03 @AquinF03 simply the best.

[Neha Sarda @nsardaa]

every time an LLM hallucinates in production, someone says 'we should fine-tune it.'

that's not engineering, it's literally superstition with a GPU bill.

you're hiding bugs you tested for, not fixing them.

@AquinF03 is building the alternative.

[Neha Sarda @nsardaa]

@Ashf03 @AquinF03 @AquinF03 all the way!!

[Neha Sarda @nsardaa]

nobody wants to say it so i will:

we're deploying AI into hospitals, banks and cars

and we genuinely don't fully know what these models learned

interpretability is the field trying to fix that

@AquinF03 is building for it

[Ash @Ashf03]

Fine-tuning can become a unsecured attack surface!

Anyone who can touch training data can attack your model and nobody's talking about this.

Ergo, At @AquinF03, we built the strongest security check systems!

Here's how it works:

[Neha Sarda @nsardaa]

compiler changed software, with it you could trace bugs to source.

AI has no equivalent. when models hallucinate, there's no way to debug. just retrain & hope.

@AquinF03 built the debugger for models, tracing exactly what broke. if you're developing AI/ML, this is for you!