Ali Khosh

Emerging tech executive, currently building ML products @meta | @pytorch. Part-time faculty at @USCGouldLaw. metalhead.

[Ali Khosh @tweetingnonstop]

These two photos are taken 13 years apart.
Top: Easter Parade on 5th Avenue, New York (1900) – Can you spot the car?
Bottom: Easter Parade on 5th Avenue, New York (1913) – Can you spot the horse?
This is what #disruption looks like.

[Ali Khosh @tweetingnonstop]

@JoytheApp Hi.
However, we cannot retrieve our hotel reservation when we click on the "Manage Booking" link in our conf email, or by calling Hyatt, or via your website. We need to make a change to this reservation by updating the dates. Please advise.

[Software Mansion @swmansion]

We're excited to announce react-native-executorch v0.3.0 packed with powerful AI features 🎉

What's New:
💬Speech to Text transcription using Whisper and Moonshine
📙OCR for extracting text from images
🚀 MPS Support
⚡️ Hookless API
🧠 LLM Message History
⬇️ Background Model Downloads
📊 Benchmarks
🧪 Experimental OCR for extracting vertical text from images

[Ali Khosh @tweetingnonstop]

Super proud of being part of this community! To many more of these in 2025 🤘🔥🚀

[PyTorch @PyTorch]

🌎 Reflecting on 2024: A year of incredible PyTorch community events!

From meetups in Seattle, Shanghai, and Korea to global hackathons, webinars and the PyTorch Conference 2024, we celebrated learning, collaboration and innovation in AI.

➡️ Highlights and what's ahead in 2025: https://t.co/QiNC0ZWJ1J

#pytorch #opensource #ai

[Ali Khosh @tweetingnonstop]

The @LinkedIn iOS app #ux is so painfully bad.

The feed rotates when the phone rotates, it shouldn’t.

The native in feed videos don’t rotate, they should!

And for every single video I’ve watched the past several months, it tells me I can swipe up. I swear I “Got It”!

[Ali Khosh @tweetingnonstop]

Great work by @swmansion with their React Native ExecuTorch library. You can easily integrate an LLM into your react native apps and run it offline!

[Software Mansion @swmansion]

What if we told you LLaMA 3.2 runs offline in @reactnative with just a few lines of code? 📲🦙

AI integration has never been this easy! Check out the video about our latest, ExecuTorch-powered React Native library 👀

https://t.co/t98ZZpbhgD

[Ali Khosh @tweetingnonstop]

🤘🙏🤘

[Francesco Capuano @_fracapuano]

Loving talking with people from @PyTorch @tweetingnonstop -- I love so much what you have done with the best framework out there 🙌🏼

[Ali Khosh @tweetingnonstop]

120+ people across 22 teams are hacking for the next 2 days building the future of AI-powered on-device solutions based on open-source models and tools from @meta and @huggingface.

Huge thanks to our friends at @join_ef @Scaleway and @ekimetrics for helping us organize this🤘

[Software Mansion @swmansion]

We've been actively exploring AI on mobile devices & we are excited to introduce React Native ExecuTorch 🚀!

A new library that enables the execution of AI models with just a few lines of JavaScript, without the need for API calls or external infrastructure. 🤩

More in 🧵👇

[Ali Khosh @tweetingnonstop]

@RezOtaku Glad to see this work Barry. I'm the PM with the ExecuTorch team. Feel free to DM me if you have any comment/faced any issued.

[Ali Khosh @tweetingnonstop]

The biggest, most participated @pytorch Conference wrapped up a few days ago in San Francisco talking
🧠 LLMs
💻 Open Source AI
📱Edge Computing

See y'all next year! 🤘
#AI #DeepLearning #PyTorchConf #OpenSource #ML #PyTorch
https://t.co/JD0CZ9um0a

[Ali Khosh @tweetingnonstop]

@airfrance I've spent north of 6 hours on calls, and WhastApp to book a seat with bassinet and a regular seat next to it. I am not getting anywhere. Is there a tier 2 support available via X?

[Ali Khosh @tweetingnonstop]

Do Barbie dolls have LLMs yet?

[Soumith Chintala @soumithchintala]

A long needed PyTorch example --
export models to desktop and mobile apps (without python), providing a native alternative to llama.cpp and other great projects in this space.
The exported models are compiled (using torch.compile) and are competitive on speed and memory use.

[Ali Khosh @tweetingnonstop]

We just released torchchat: Runnig LLMs on mobile, desktop and laptop with no connection needed, more private, and more personalized.🔥🤘🚀 #LLMs #ondeviceAI @PyTorch
https://t.co/lCeM8ZTvDI

[Software Mansion @swmansion]

Here comes our fresh #ExecuTorch tutorial by Wojtek Jasiński✨showing how to use the @PyTorch-based framework to export and run a style transfer model in your mobile apps on Android 📲

Get started ➡️ https://t.co/0XA1qZGMux

#ReactNative #PyTorch #ExecuTorch #AI

[Ali Khosh @tweetingnonstop]

Thank you for doing this. I always wondered how the story would end.

[Pierrick Chevallier | IA @CharaspowerAI]

I've been wanting to get to the bottom of this story for so long. #DreamMachine #LumaAI 😂

[Ali Khosh @tweetingnonstop]

@Lee_Jeffries @Metallica Happy BD Lee! What a killer shot. It so well captures what this band is all about!

[Ali Khosh @tweetingnonstop]

Just released ExecuTorch alpha, which brings LLMs to mobile, desktop & other edge devices. You can run Llama2 7B on Galaxy S24 at ~11 tokens/sec, fully on-device. And it also supports running Llama 3 8B on mobile phones! We're so on the edge right now. 🚀 https://t.co/R6C6k7PUda

[Andrew Ng @AndrewYNg]

Meta released Llama 3 on my birthday! 🎂 Best present ever, thanks Meta! 😀

[Ali Khosh @tweetingnonstop]

@nikitabier Manhattan Beach beats that neighborhood in almost every single aspect IMO.