Camilla Montonen
@spimescape.bsky.social
Building recommender systems @ Consumer Tech Co
This is how I learnt about GPUs without basically any background knowledge.
this is the way to learn with ai. you can start anywhere and backfill. you have to remain curious and careful and not just passively eat up plausible explanations. but if you put in the effort and the model is good, it’s powerful
The old path: Learn Lean (6 weeks), study abstract algebra (8 weeks), understand group theory (4 weeks), finally attempt your proof.
The new path: Start with your exact problem. Generate a tutorial for it. Backfill concepts as you hit them.
The new path: Start with your exact problem. Generate a tutorial for it. Backfill concepts as you hit them.
October 20, 2025 at 7:12 PM
This is how I learnt about GPUs without basically any background knowledge.
Reposted by Camilla Montonen
modern LLM inference engines like vLLM & SGlang are becoming tough to dive into. to learn how these inference engines work, nano-vllm is a fantastic educational project—complete Page Attention & LLM scheduler in <1k loc.🤯
flaneur2020.github.io/posts/2025-1...
flaneur2020.github.io/posts/2025-1...
A Walkthrough of nano-vllm | Flaneur2020
Recently, I’ve been delving into the architecture of production-grade inference engines. While projects like vLLM and SGLang are crazy sophisticated, …
flaneur2020.github.io
October 12, 2025 at 3:43 PM
modern LLM inference engines like vLLM & SGlang are becoming tough to dive into. to learn how these inference engines work, nano-vllm is a fantastic educational project—complete Page Attention & LLM scheduler in <1k loc.🤯
flaneur2020.github.io/posts/2025-1...
flaneur2020.github.io/posts/2025-1...
TIL that junk journaling is a thing!
October 12, 2025 at 11:24 AM
TIL that junk journaling is a thing!
Microblogging like it's 2010s with vibes.
October 12, 2025 at 9:06 AM
Microblogging like it's 2010s with vibes.
Me after hearing that Python 3.14 has removed the GIL:
"ah, finally increased throughput of pulling data from Big Query into my Jupyter notebooks."
Also me: "ah, a new footgun to add to my repertoire"
"ah, finally increased throughput of pulling data from Big Query into my Jupyter notebooks."
Also me: "ah, a new footgun to add to my repertoire"
October 11, 2025 at 5:55 PM
Me after hearing that Python 3.14 has removed the GIL:
"ah, finally increased throughput of pulling data from Big Query into my Jupyter notebooks."
Also me: "ah, a new footgun to add to my repertoire"
"ah, finally increased throughput of pulling data from Big Query into my Jupyter notebooks."
Also me: "ah, a new footgun to add to my repertoire"
One of the design issues with Jupyter notebooks when it comes to heavy ML workloads is that the notebook server runs by default on the same machine as the kernel that executes the code.
October 11, 2025 at 5:53 PM
One of the design issues with Jupyter notebooks when it comes to heavy ML workloads is that the notebook server runs by default on the same machine as the kernel that executes the code.
NVIDIA seems to invest a lot of engineering effort into making higher level libraries for writing efficient GPU code and yet everyone is flexing by rolling out their own CUDA kernels.
October 11, 2025 at 3:13 PM
NVIDIA seems to invest a lot of engineering effort into making higher level libraries for writing efficient GPU code and yet everyone is flexing by rolling out their own CUDA kernels.
Reposted by Camilla Montonen
A speech about what drives me, how science and open source are bitter victories, unable to make improve the world if society does not embrace them for the better:
gael-varoquaux.info/personnal/a-...
gael-varoquaux.info/personnal/a-...
A national recognition; but science and open source are bitter victories
I have recently been awarded France’s national order of merit, for my career, in science, in open source, and around AI. The speech that I gave carries messages important to me (French below;...
gael-varoquaux.info
October 10, 2025 at 11:37 AM
A speech about what drives me, how science and open source are bitter victories, unable to make improve the world if society does not embrace them for the better:
gael-varoquaux.info/personnal/a-...
gael-varoquaux.info/personnal/a-...
I was looking for a solution to "migrate a container that is close to OOM" onto another node and found CRIU.
Still a bit unclear if it supported on Google's GKE or not.
Still a bit unclear if it supported on Google's GKE or not.
October 11, 2025 at 1:31 PM
I was looking for a solution to "migrate a container that is close to OOM" onto another node and found CRIU.
Still a bit unclear if it supported on Google's GKE or not.
Still a bit unclear if it supported on Google's GKE or not.
Container image experts - is it possible to manually create a new layer by manipulation the files in the tar archive you get after running docker image save?
October 8, 2025 at 8:17 PM
Container image experts - is it possible to manually create a new layer by manipulation the files in the tar archive you get after running docker image save?
One of the major design flaws of many notebook environments like Jupyter is that the kernel that does computations is not separate from the machine that runs the notebook server itself.
October 8, 2025 at 8:15 PM
One of the major design flaws of many notebook environments like Jupyter is that the kernel that does computations is not separate from the machine that runs the notebook server itself.
When folks say they are going to build AGI - what exactly does that look like?
October 5, 2025 at 2:55 PM
When folks say they are going to build AGI - what exactly does that look like?
I really miss London.
September 30, 2025 at 8:51 PM
I really miss London.
I am very much a beginner in this space but napkin math for systems problems is quite a lot of fun!
September 30, 2025 at 7:53 PM
I am very much a beginner in this space but napkin math for systems problems is quite a lot of fun!
I use k8s extensively in the AI/ML space and would love to see a k8s that is more geared towards ML pipeline or batch job needs.
The reason we stick with k8s despite the shortcomings is the sheet number of cloud services and oss platforms that integrate with k8s out of the box.
The reason we stick with k8s despite the shortcomings is the sheet number of cloud services and oss platforms that integrate with k8s out of the box.
so long! god, VMWare got 15+ years before k8s came around, and it's barely been 10 for k8s
Seriously though, there needs to be a new theory of what is interesting/important, and the only space doing that is AI-related stuff where they're mostly using even older tech (slurm)
Seriously though, there needs to be a new theory of what is interesting/important, and the only space doing that is AI-related stuff where they're mostly using even older tech (slurm)
September 30, 2025 at 7:50 PM
I use k8s extensively in the AI/ML space and would love to see a k8s that is more geared towards ML pipeline or batch job needs.
The reason we stick with k8s despite the shortcomings is the sheet number of cloud services and oss platforms that integrate with k8s out of the box.
The reason we stick with k8s despite the shortcomings is the sheet number of cloud services and oss platforms that integrate with k8s out of the box.
Interesting but it looks like it also targets services so perhaps not so suitable for ML batch jobs.
"Running a shared-nothing architecture at the edge, we needed a simple way to scale HTTP/TCP based containers without the overhead of complex infrastructure or additional dependencies. "
"Running a shared-nothing architecture at the edge, we needed a simple way to scale HTTP/TCP based containers without the overhead of complex infrastructure or additional dependencies. "
Just came across this, looks interesting medium.com/air-pipe-blo.... Uses @cloudflare.social Pingora
Orbit: A Lightweight Container Orchestrator Built in Rust
Discover how we’re building a lightweight ~5MB alternative to container orchestration, using Rust and Cloudflare’s Pingora framework.
medium.com
September 30, 2025 at 7:43 PM
Interesting but it looks like it also targets services so perhaps not so suitable for ML batch jobs.
"Running a shared-nothing architecture at the edge, we needed a simple way to scale HTTP/TCP based containers without the overhead of complex infrastructure or additional dependencies. "
"Running a shared-nothing architecture at the edge, we needed a simple way to scale HTTP/TCP based containers without the overhead of complex infrastructure or additional dependencies. "
Reposted by Camilla Montonen
Is anyone working on a K8s alternative? Kinda curious.
June 24, 2025 at 5:52 PM
Is anyone working on a K8s alternative? Kinda curious.
What is the best place to ask noob questions related to the Linux kernel?
September 30, 2025 at 7:05 PM
What is the best place to ask noob questions related to the Linux kernel?
Any webmasters here? I recently discovered that the crawler with the user agent ChatGPT-User does not appear to respect robots.txt, but it also could be that I just misconfigured my robots.txt. Anyone else had a similar experience?
September 21, 2025 at 9:16 AM
Any webmasters here? I recently discovered that the crawler with the user agent ChatGPT-User does not appear to respect robots.txt, but it also could be that I just misconfigured my robots.txt. Anyone else had a similar experience?
Are there are any k8s community initiatives that aim to improve container cold start times on clusters?
September 17, 2025 at 3:54 PM
Are there are any k8s community initiatives that aim to improve container cold start times on clusters?
Computer vision folks, what is the current SOTA when it comes to circle detection in images?
September 14, 2025 at 12:44 PM
Computer vision folks, what is the current SOTA when it comes to circle detection in images?
I thought that the ultimate winners of the current era would be the chip manufacturers and the infra software layers, but this gave me some new perspectives to think about.
AI Will Not Make You Rich, by @ganeumann.bsky.social
I find this essay extremely compelling. Especially the points made about who got rich from shipping containers.
joincolossus.com/article/ai-w...
I find this essay extremely compelling. Especially the points made about who got rich from shipping containers.
joincolossus.com/article/ai-w...
AI Will Not Make You Rich
The disruption is real. It's also predictable.
joincolossus.com
September 14, 2025 at 10:04 AM
I thought that the ultimate winners of the current era would be the chip manufacturers and the infra software layers, but this gave me some new perspectives to think about.
London friends, I hope you are staying safe.
September 14, 2025 at 9:38 AM
London friends, I hope you are staying safe.
Reposted by Camilla Montonen
Kindness is punk as fuck.
September 14, 2025 at 2:30 AM
Kindness is punk as fuck.
Makes me sad that Glitch was shutdown!
September 13, 2025 at 7:42 PM
Makes me sad that Glitch was shutdown!