Ljubomir Josifovski

ljupco

https://ljubomirj.github.io/

AI & ML interests

Now - ML/AI, agents, auto-science & engineering, forecasting. Previous - systematic trading, research & development. Previous^2 - speech recognition in noise, speech synthesis, machine learning.

Recent Activity

liked a model about 23 hours ago

akhilaaa3/Jev-Omni

liked a model about 23 hours ago

chinhnc/Quyet-1.0-Large

liked a model about 23 hours ago

Remek/basal-omni-1.1-max

View all activity

Organizations

None yet

upvoted a collection about 23 hours ago

basal-omni-1.0

Collection

Typed decisions over text, images and audio, served by the basal engine v2.0.0 (github.com/rkinas/basal). • 4 items • Updated 2 days ago • 5

upvoted an article 2 days ago

Article

The Most Intuitive Way to Understand Score Function Centering: Understand Trainer-Generator Mismatch in Async RL from the Score Function

NormalUhr

•

4 days ago

• 7

upvoted an article 5 days ago

Article

Multimodal open d1 decision models for the edge

LiquidAI

•

12 days ago

• 40

upvoted a collection 6 days ago

Amos 1

Collection

Open-weight decision models from 0.5B to 27B: typed questions in, a probability for every answer out. Apache 2.0. • 4 items • Updated 6 days ago • 12

upvoted a paper 16 days ago

DeepSeek-V4.1-Flash: Pushing the Limits of KV Cache Compression

Paper • 2609.19969 • Published Sep 17 • 230

upvoted a collection 28 days ago

MiMo-V2.6

Collection

6 items • Updated about 5 hours ago • 134

upvoted 3 collections about 1 month ago

upvoted an article about 1 month ago

Article

Keep the Tokens Flowing: Lessons from 16 Open-Source RL Libraries

aminediroHF, qgallouedec, kashif, lewtun, edbeeching, albertvillanova, nouamanetazi, lvwerra, sergiopaniego

•

Mar 10

• 195

upvoted a paper about 2 months ago

Language Models Can Control Their Own Attention

Paper • 2609.02737 • Published Sep 2 • 80

upvoted 4 collections about 2 months ago

Spark-X2.5

Collection

Spark-X2.5 is a compact, general-purpose language model for conversation, writing, translation, reasoning, coding, tool use, and agentic workflows. • 10 items • Updated Sep 16 • 50

Qwen3-ASR

Collection

7 items • Updated Jun 26 • 84

Technetium GA-II GenerativeAI Design Sets

Collection

De novo AI-designed small-molecule ligands docked into therapeutic targets, as protein-ligand complexes (Technetium GA-II platform). • 26 items • Updated 1 day ago • 1

Qwen3.8 27B Uncensored / Abliterated

Collection

The Qwen3.8-27B-Uncensored family from OrcaRouter 🐋 FP8 ✅ NVFP4 ✅ GGUF ✅ INT8 ✅ MLX ✅ Full precision ✅ • 6 items • Updated Aug 20 • 112

upvoted a paper 2 months ago

Agent Lightning v1.0: Towards Harnessed Agentic RL

Paper • 2608.17528 • Published Aug 18 • 35

upvoted 2 collections 2 months ago

Qwen 3.8 27B

Collection

11 items • Updated 3 days ago • 4

dots.llm1

Collection

2 items • Updated Jun 11, 2025 • 19

upvoted an article 2 months ago

Article

Lattice: an 8 MB static retriever that embeds Wikipedia in 7 minutes

erikkaum

•

Aug 7

• 20

upvoted a paper 3 months ago

AskChem: Claim-Centered Infrastructure for Chemistry Literature Synthesis

Paper • 2607.28618 • Published Jul 30 • 310

Ljubomir Josifovski

AI & ML interests

Recent Activity

Organizations

ljupco's activity

The Most Intuitive Way to Understand Score Function Centering: Understand Trainer-Generator Mismatch in Async RL from the Score Function

Multimodal open d1 decision models for the edge

Keep the Tokens Flowing: Lessons from 16 Open-Source RL Libraries

Lattice: an 8 MB static retriever that embeds Wikipedia in 7 minutes