ML Researcher | Core contributor to Apples official ML framework (MLX)| Violin enthusiast 🎻 | Always coding, sometimes watching Anime.
Germany
Joined February 2022
- Tweets2.3K
- Following894
- Followers3.2K
- Likes70K
Pinned Tweet
My name got featured in one of Apples WWDC 2025 session videos 👀🍎
Crazy to think my work made it into a global Apple dev session. I’m proud!!
Thanks @awnihannun @angeloskath @Apple
OMG someone already stated a PR 😱😱😳
github.com/Neywa-Labs/mlx-vl…
Had to file a issue to @Neywalabs MLX-VLM, we need ASR training!!!!!
github.com/Neywa-Labs/mlx-vl…
1
3
568
Had to file a issue to @Neywalabs MLX-VLM, we need ASR training!!!!!
github.com/Neywa-Labs/mlx-vl…
The MacOS native JOSIE App will be released on 01.December!!!!!
JOSIE: A local, real-time, full-duplex voice and chat assistant that listens, knows who's speaking, talks back with a natural voice, and takes real action in your Reminders, Calendar, Maps and Messages.
goekdeniz-guelmez.github.io/…
I might have to rename it, since that demo video I have added a lot more agentic things. I thing I should rename “JOSIE-Voice” to just “JOSIE”.
Here is a 6 minute call within JOSIE-Voice on a MacBook Pro M4 Pro using: Qwen-3 0.6B TTS/ASR 4bit and Deepseek 4.1 Flash on ollama cloud. it does 100+ Tool calls with agentic work in the real world. No cuts or speedup's.
1
11
873
Had to buy me an Apple Studio Display 5K, can’t wait to use that, my current one makes my eyes hurt.
2
6
741
Can do the same on Apple silicon now too 😎😎 with a very cool real time view and a chat window to ask codex or Claus directly ✌️
github.com/Goekdeniz-Guelmez…
16
1,046
Trying out @liquidai LFM3-Omni on JOSIE-Voice, it’s actually pretty cool and almost has a very uncanny voice. The cool thing is that it has native tool calling so like OpenAI GOT-Live-1 and Moshi RAG you can have synchronized tool delegation. Core backend used is mlx-audio-swift by @Neywalabs still has some bugs though the model sometimes generates only text without the audio, that was the error you see in the end. I used the full precision bf16 model and it was pretty fast on my M4 Pro
Lets keep the releases coming!
Introducing MLX LoRA Studio 3.0.0 🚀🚀🚀
MLX LoRA Studio 3.0.0 brings two ned training algorithms DSLA and KLPO, and finally brings the whole trainer to your favourite agent's fingertips, like @Teknium Herms Agent, @OpenAI Codex @AnthropicAI Claude Code, and al the other that support Skills and MCP with a one click installer.
Whats also new:
- Multiple queued training runs.
- Automatic run cancelation when loss becomes "nan".
- A new live MCP server dashboard
- And more speed and recourse usage optimisations.
Download and try it out:
goekdeniz-guelmez.github.io/…
🚀🚀🚀
Today we're coming out of stealth, introducing Neywa Labs 🚀
We built the infrastructure to build and deploy AI that runs on the devices people already own. Private, offline, unlimited.
One mission: make AI available to everyone in the world.
neywalabs.ai/
16
1,319
MLX-VLM's trainer now supports MTP 🚀🚀🚀
github.com/Blaizzy/mlx-vlm/p…
Finishing up the last two PRs for base VLM training for MLX-VLM by @Prince_Canuma to then start adding decision model training. And then………
Here are the current PRs:
Multi-Token-Prediction: github.com/Blaizzy/mlx-vlm/p…
Quantization Aware Training: github.com/Blaizzy/mlx-vlm/p…
Open questions I have for the new Full Duplex Dessa model:
- Is it only pertained on audio? If so, then how can we fine-tune it? Or will there be fine-tuned models released along side it?
- How can we add async tool calling into it when it was trained on audio only?
5
15
2,643
Just tell Codex to train a custom model for you on apple silicon!!!!!
Here is a real example of codex pre-training a tiny LLM and iterating on it to get the best results, using MLX-LM-LoRA's MCP part.
How to reproduce it:
1. Install mlx-lm-lora: python3 -m pip install -U "mlx-lm-lora[mcp]"
2. Install the skill: npx skills add Goekdeniz-Guelmez/mlx-lm-lora --skill mlx-lm-lora -a codex -g
3. Find out what path the mcp package is: which mlx-lm-lora-mcp
4. Copy the path and copy it into here: codex mcp add mlx-lm-lora -- {path}
5. Start chatting
Here is the prompt I used:
"""Task: Find the best pre-training run for tiny-random/qwen3
Use the mlx-lm-lora MCP server and skills only. No other training stack.
Fixed
Model: tiny-random/qwen3 (if it doesn't resolve, ask me before continuing)
Dataset: mlx-community/recycling_the_web-1k
Task: pre-training
Steps: 100
Global batch size: 8
Context length: 128
Do
Run a single SFT pre-training recipe 4 times, letting you pick the hyperparameters and optimizer trial after trial yourself. Keep the fixed constraints identical every run. Use a different seed per trial and evaluate each on a held-out slice.
Deliver
A table of the 4 trials: optimizer, lr, schedule, seed, final train loss, val loss/perplexity.
The best pre-training trial by val loss, with its config and saved adapter path.
If any fixed constraint can't be satisfied by the MCP server, stop and tell me."""
I've gotten a lot of comments saying it's just my FYP algorithm feeding me more voice AI content, but I don't think that's true here. Realistically is probably a massive coincidence, with all of us working toward similar goals at the same time.
But just to look at this timeline:
October 18 - First JOSIE-Voice teaser video
October 20 - Mistral Voxtral 3 + ElevenLabs Eleven Realtime v1 + Kyutai Pocket TTS v4.0.0
October 21 - Gradium Phonon TTS release
October 22 - IBM Granite Speech 5.0 + Amazon Polly Generative
October 23 - Block Dessa full-duplex voice model teaser
Within the same week as my JOSIE-Voice teaser, multiple teams released technology directly relevant to building efficient, real-time, local, voice agents. I'm not saying there's a conspiracy here. I'm just saying the timing is getting suspiciously cinematic. 🤔
jk, its probably just coincidence 😂😂😂