Jude Koome

@ilyasut layed the blueprint

[Jude Koome @koomejude]

It was an amazing use of Graph Databases ( Neo4j) and we intend to continue using the Technology in our solutions and workflows

[Kenya AI Challenge @KenyaAIHack]

🌾 Congrats to ShambaLadder!

@MC_AgriFin Track Finalists at the Kenya AI Challenge. 🏆

Built by @koomejude , Collins Mutinda, Mark Mulei, John Nyajowi & Jim Dave.

🌐 https://t.co/gIimnIHQWF

@DirectEdDev @AiKenya1 @Lovable @neo4j @FeatherlessAI

[Jude Koome @koomejude]

@shownotover Congratulations 🎉 man

[Sebastian Raschka @rasbt]

Just saw that the LLMs-from-scratch repository passed 100,000 stars on GitHub!
This is super cool and motivating. I am really happy to see that this open-source repo has helped so many people.
Thanks also to everyone who shared ideas and opened PRs with improvements!

Of course, I plan to keep adding new material, including new attention variants and architectures (while bigger projects like RL and Reasoning From Scratch live in their separate repositories).

I am also currently working on a larger applied custom “small” LLM project. It has been keeping me super busy this month, but I will share more on that soon.

If you are new to it, some of the highlights include

1. Of course, the complete code path from tokenization and attention to pretraining, classification, and instruction fine-tuning, etc. All of it FROM SCRATCH, of course! (RL lives in a companion repo.)

2. From-scratch implementations of Llama, Qwen, Gemma, and Olmo (smaller variants that run locally and can be plugged into the training scripts).

3. From-scratch implementations of attention alternatives and other architecture components, such as GQA, MLA, sliding-window attention, Gated DeltaNet, DeepSeek Sparse Attention, cross-layer KV sharing, and mixture-of-experts

4. Materials on KV caching, training performance, memory-efficient weight loading, DPO, evaluation, and LoRA

So, if you don’t have any weekend plans yet, happy tinkering!

[Jude Koome @koomejude]

More like 2019

[Jude Koome @koomejude]

@NeerajGoesAi Hello Neeraj tried to talk to the voice assistant but it's failed due to trying to wake up. Very excited to see your portfolio. Hope you get that fixed soon. To give us the full experience

[Jude Koome @koomejude]

@NeerajGoesAi @SakshiSugandhi Which type and secs

[Jude Koome @koomejude]

@waterloo_intern I support hard work but you're making the life expectancy rates make sense?

[Kenya AI Challenge @KenyaAIHack]

🥈 Congrats to Angawatch!

Runner up in the @MC_AgriFin and @neo4j Tracks, and @MasumiNetwork Track Winner at the Kenya AI Challenge.

Built by Ryan Nguitui & Gitahi Githinji.
🌐 https://t.co/q41qcXgkdn

@DirectEdDev @AiKenya1 @MasumiNetwork @Lovable @neo4j @FeatherlessAI

[Kenya AI Challenge @KenyaAIHack]

🌾 Congrats to ShambaLadder!

@MC_AgriFin Track Finalists at the Kenya AI Challenge. 🏆

Built by @koomejude , Collins Mutinda, Mark Mulei, John Nyajowi & Jim Dave.

🌐 https://t.co/gIimnIHQWF

@DirectEdDev @AiKenya1 @Lovable @neo4j @FeatherlessAI

[Jude Koome @koomejude]

Kenyan highschools need this but problem is we play football in class

[iPaul @iPaulCanada]

2026世界人工智能大会，某科技公司展示的跨时代教具--AI黑板

[Swati Gupta @hrswatigupta]

Anthropic engineer:

"You're not supposed to prompt Claude. You're supposed to build a system that prompts itself."

In 45 minutes, she breaks down how Anthropic builds agents that remember, learn from their mistakes, and get smarter with every run.

Worth more than any paid course you'll find on building agents.

Watch this and bookmark

[Codez @0xCodez]

Andrew Ng:

“AI agents are doing almost 100% of my tasks now - the hype has exceeded my expectations.

in 3-6 months, we’ll all be building graphs to orchestrate self-improving agents. No more prompting.”

In a 40-minute talk, Andrew Ng explains how to build self-improving agentic systems from scratch.

Worth more than a $500 agentic course.

Watch this video, then read the article below on how to become a graph architect.

[Jude Koome @koomejude]

Finally an amazing read on Low resource Kenyan languages

[C-elo Labs @c_elolabs]

“Low-resource” does not simply mean “too little audio.”

From a Multilingual Streaming ASR Backbone to Kenyan-Language Systems: Data-Centric Adaptation of Nemotron 3.5 for Kikuyu, Dholuo, and Kalenjin

The paper: https://t.co/0wOZ1by8SG

@NVIDIAAI

[Khairallah AL-Awady @eng_khairallah1]

this is f*cking gold

Andrej Karpathy joined Anthropic five weeks ago.

A friend on his team just showed me the exact LOOPS.md file he actually uses.

I dropped it into my setup. The very first response was different.

Not slightly different. Completely different.

Claude stopped giving generic answers and started working exactly the way I think.

You don't talk to the model anymore. You build the system that talks to the model for you.

Bookmark it before it gets lost in your feed.

Read it now, then check the article below.

[Jude Koome @koomejude]

@iam_chonchol Drive😞

[Jude Koome @koomejude]

@iam_chonchol Me here on 10th July🙂🙏🏾🙏🏾

[Kasif @md_kasif_uddin]

Day 2 of Learning C++ 🥲

[Swati Gupta @hrswatigupta]

Anthropic pays $750,000+ a year for engineers who can build LLM architectures from scratch. Stanford taught the entire thing in 1 hour lecture & released it for free.

Bookmark & watch this today before someone takes it down and read this article below

[Mark A100 @gatere_mark]

“a billion parameters of local data”? 👀 Brooo, this phrasing doesn’t even make sense technically.

Secondly, 78% “accuracy” on local dialects - what metric? WER? Custom test set? Without details or reproducible evals, this just sounds like classic hype.

“train, re-train, extra training, during training, 'training the model in it's initial stages' …”
Training even a 1B model from scratch requires way more than a single GPU - distributed training, for something like TinyLlama 1.1B took about 16 A100-40G GPUs for a span of 90 days. I hope he meant fine-tuning, which I also highly doubt.

"42 Languages" 🙄 ... No technical details provided - no architecture, no data curation method, no training process, compute used or reproducible evals.

Did you also “train the model” to write Rust and Golang? Convince me this isn’t just a GPT wrapper.

I support the effort, but some things need to be called out for clarity.

[Reethu @ritu_twts]

Be honest devs, Is coding still worth learning in the AI era?