ML Engineering @adobe prev: @huggingface
Joined January 2009
- Tweets 2,456
- Following 1,197
- Followers 6,813
- Likes 8,235
356 Photos and videos
Apr 29
Asked TALKIE-1930 what 2026 would look like
Very cool model btw!
Apr 27
Announcing Talkie: a new, open-weight historical LLM! We trained and finetuned a 13B model on a newly-curated dataset of only pre-1930 data. Try it below!
with @AlecRad and @status_effects 🧵
1
412
Mar 12
FLUX.2 [klein] 9B KV cache optimization. This is going to be fun
Mar 12
My favorite editing model, FLUX.2 [klein] 9B, just got 2x faster:
Meet FLUX.2 [klein] 9B-KV 😍💨
> Using KV-Cache Optimization to reduce computation & speed up inference by up to 2.5 times for multi-reference editing
love how well it edits "around" the bullets
4
660
6 Nov 2025
Saw this random video on my timeline and I really wished it ended like this
3
679
29 Oct 2025
testing fofr benchtest with new Firefly Image 5
"an upside down giraffe"
2
18
4,609
28 Oct 2025
before a live demo: GPT god, please follow my instructions 🙏
1
358
30 Sep 2025
Testing nano-banana with recursive editing from simple prompts
"Show this scene one second in the future"
90
109
2,730
375,671
30 Sep 2025
inspired by x.com/genekogan/status/19161…
"Subtly alter this image by replacing one or two details with something unexpected and funny"
26 Apr 2025
Another gpt-image feedback loop: "Generate the same photo 5 seconds into the future."
Starting with distracted boyfriend meme.
1
4
79
15,635
30 Sep 2025
Thanks @fal for the credits and easy API to switch models! Script more examples here:
github.com/radames/nano-bana…
24
9,240
22 Sep 2025
Fascinating insights from Surge AI CEO @echen on the data generation business
youtube.com/watch?v=X39OZndI…
Had a great conversation with @l2k on Gradient Dissent about what's actually happening in post-training right now.
We dug into why LMSYS is setting the industry back - researchers are literally being forced to make models hallucinate more to climb the leaderboard...
1
891
25 Jun 2025
very cool idea to expose Claude API inside an artifact
await window.claude.complete("Hello, how are you?")
25 Jun 2025
Introducing two new ways to create with Claude:
A dedicated space for building, hosting, and sharing artifacts, and the ability to embed AI capabilities directly into your creations.
3
729
18 Jun 2025
People were having fun!
12 Jun 2025
My project Shape Shifter is part of @exploratorium new Adventures in AI exhibition! Real-time image diffusion for kids and adults. Opens today in San Francisco with special After Dark event tonight!
6
698
12 Jun 2025
My project Shape Shifter is part of @exploratorium new Adventures in AI exhibition! Real-time image diffusion for kids and adults. Opens today in San Francisco with special After Dark event tonight!
2
2
14
2,784
12 Jun 2025
This project is based on my open source work exploring real-time capabilities of diffusion models.
github.com/radames/Real-Time…
3
304
27 Apr 2025
oww @Gradio now has an official fancy image slider component. I upgraded my old Enhance This Space with it!
5
10
118
20,823