Developer Relations @Datadoghq. Previously @AWS, @Clarifai, @Auth0, @SendGrid Opinions are my own.
New York, NY
Joined March 2009
- Tweets2.7K
- Following2.6K
- Followers1.8K
- Likes8.2K
👀
Cool new agent skill for agent assisted conversion of transformers models to MLX! 🔥
Blog post by @awnihannun and @pcuenq
huggingface.co/blog/transfor…
1
2
257
Been working on this for a bit, stoked to announce a community port of the AWS Strands Agents framework to Swift. It can run natively on device and you can build those fun OpenClaw-style computer use examples. It can run inference locally via MLX or through Bedrock.
In the video I connected an agent to a Mac MCP server and asked it to open Word and write a sentence. It also sends OTel telemetry to your tool of choice. In the video you can see the telemetry from the agent being shown in Datadog's LLM Observability.
communitystrands.com/swift
Coming soon: Interacting with the agents through voice either AWS Nova Sonic or locally with @Prince_Canuma's mlx-audio-swift package.
cc: @awsdevelopers, @awnihannun , @shshnkp
3
6
359
Introducing MLX-Swift-TS
github.com/kunal732/MLX-Swif…
An SDK for running time series foundation models fully on-device on Apple Silicon.
When I joined @datadoghq , I was introduced to Toto, our time series foundation model, and got excited about zero-shot forecasting across different domains. While building a health copilot app, I realized there wasn’t a simple way to run models like these locally on device. So I built one.
MLX-Swift-TS exposes a common TimeSeriesForecaster interface for loading and running multiple time series architectures directly in Swift using MLX. No server required.
The attached video shows on-device forecasting running inside a native Swift app.
Huge thanks to @awnihannun and the MLX team for building MLX and its Swift API, @Prince_Canuma for inspiration on MLX SDK patterns, and @atalwalkar and the Datadog team for Toto.
@Prince_Canuma Do you have or know any resources to help convert times series foundation models like google's timesfm to MLX ? From my understanding these use patch embeddings for inputs in a similar way to how image inputs are encoded for vlms .
1
46
Just converted @arcee_ai 's Virtuoso Lite model to MLX. A 10B parameter model distilled from DeepSeek V3. Have been impressed with it's performance. Recommend watching @julsimon's video to learn more youtube.com/watch?v=nrqJb1I7…)
If you want to deploy it in your Swift app - you can check out my demo app showcasing this model running locally.
github.com/kunal732/mlx-swif…
MLX Model here: huggingface.co/mlx-community…
cc @Prince_Canuma - not sure if you were involved with this but know you're at @arcee_ai as well.
1
6
258
Running DeepSeek (huggingface.co/mlx-community…) with MLX locally in a swift application. Xcode Project can be downloaded here: github.com/kunal732/mlx-swif…
Just converted @JinaAI_ 's new ReaderLM V2 model into MLX. Trained to convert different formats like html into markdown.
MLX port over here: huggingface.co/mlx-community…
Gist of swift code that crawls a website and converts to markdown here: huggingface.co/mlx-community…
Just realized posted the wrong link for the gist: gist.github.com/kunal732/05b…
6
303
🪟 MLX Model Manager now supports Microsoft's Phi-4 LLM. You can add it to your Swift app in just a couple of lines of code!!
💡 Gist: gist.github.com/kunal732/a18…
📱 Sample App: github.com/kunal732/mlx-swif…
[🎥 Attaching a demo video below!]
Have been glued to this video of my daughter’s first bike ride. Wild how three+ years flew by in a blink of an eye.
2
275
Use @Google's PaliGemma2 VLM in your macOS Swift app in about a minute with mlx-model-manager (github.com/kunal732/MLX-Mode…)
Gist here: gist.github.com/kunal732/946…
Next will add support for other MLX models created by @Prince_Canuma . What should I start with first ?
Today I'm launching a very early release of MLX Model Manager (github.com/kunal732/MLX-Mode…). A Swift package for quickly and easily adding LLM/VLMs with just a couple of lines of code into your MacOS/iOS applications for local and private inferencing.
This is built on the amazing work of the MLX team & what they have done with github.com/ml-explore/mlx-sw…. MLX Model Manager unifies the work they have done with MLXLLM as well as MLXVLM into one package with added abstractions to simplify the usage.
I've also added support for Google's new Paligemma2 Vision Model (huggingface.co/blog/paligemm…). You can see a demonstration in the video below!
Huge shoutout to @awnihannun for the creation of MLX, David Koski on leading the charge for MLXVLM and @Prince_Canuma for the creation of MLX-VLM. Three massive pillars in the MLX community among many others.
Here is an example of it running Microsft's Phi-2 LLM (huggingface.co/mlx-community…) on my iPhone 12 Pro Max.
🎉 Excited to introduce ScreenML(screenml.com) – a free mac app that automatically generates meaningful, context-aware names for your screenshots! 🖼️✨
Getting started is easy: 1️⃣ Launch the app. 2️⃣ Allow permissions to capture screenshots. 3️⃣ Hit the cmd + 0 hotkey to generate context-aware file names in an instant.
If this resonates with you, I’d love to hear your thoughts – feedback, suggestions, and anything else you want to share!
🚀 Exciting news! The Onboto App is now available for Mac users, bringing ChatGPT and Dall-E Image Generation to any application on your desktop! 🖥️ Get fast and accurate responses, trackable prompts, and seamless integration with any app. Learn more at onboto.app
Just got my copy of The Technical Content Strategy Decoded. Congrats on getting this out there @adamd
3
311
Exciting news! Created Onboto - a Mac app that lets you use ChatGPT right from your desktop. Enter prompts in any app and get back responses, without leaving your workflow. Sign up for the waitlist & see how it can boost your productivity 🚀! onboto.app
@adamd ~8yr anniversary for the email switch hitter podcast and these sweet baseball cards. When's the next collab ? vimeo.com/93305631
1
2
I’ve been very lucky to work on an amazing team here at AWS and helping to empower our devs. Starting next wk I’ll be changing roles and leading our Modern Apps DA team. If anyone wants to work with @lenadroid, or w/any of our other amazing DA teams check bit.ly/awsdevrel-jobs
What a fun week! So many customer conversations and meeting amazing coworkers in person. #reinvent21
11
