Hyunseung Kim

Researcher @Krafton_AI
PhD Candidate @KAIST_AI

[Hyunseung Kim @mynsng_]

Finding an open model that handles Korean well was tough when building PUBG Ally.

Excited to see Raon-1 mini launch with strong Korean and agentic capabilities. And it’s really fast, too!

If you’ve faced similar challenges, give it a try.

[Kangwook Lee @Kangwook_Lee]

👋 KRAFTON's Raon is entering the open-weight scene!

Meet Raon-1 mini, our first model, trained from scratch.

On par with leading models in its size class, while running faster and using less memory.

Available under the Apache 2.0 license.
https://t.co/fOFoUcfDLc

[Kangwook Lee @Kangwook_Lee]

I wrote a piece for Fast Company about games, robots, and what we're learning from working on both.

Some thoughts on how games can help us build AI for the physical world. Please check it out!

https://t.co/GX1x6RvHyb

[Kyungmin Lee @lee_kyungmin21]

We introduce FlashDexRetarget: one RL policy that retargets many human hand-object demos to dexterous robot hands.

On a 50-motion benchmark with XHand: 90% success with ~100x less compute than CHORD.

Project page: https://t.co/jrxYmM7pBA
Paper: https://t.co/wusffvWSeb

[Kyungmin Lee @lee_kyungmin21]

Excited to share that PHUMA has been accepted to #CoRL2026 as a Spotlight! 🎉

Huge thanks to my co-authors (Sibeen Kim, @leeyngdo , @mpark1999 , @mynsng_ , @dongyoon_ai, @kty2396) and to @hojoon_ai and Prof. Jaegul Choo for their guidance. See you in Texas!

[Hyunseung Kim @mynsng_]

@mkkang_1133 Thanks Minki !!!

[Hyunseung Kim @mynsng_]

Excited to share the PUBG Ally Tech Report!
📄 https://t.co/kiVncTRppn

In June, we introduced PUBG Ally, the world’s first in-game AI agent.

Check out the unique challenges behind its architecture, data, training, evaluation, and on-device optimization!

[Kangwook Lee @Kangwook_Lee]

https://t.co/nIvkmeIQFD

[Dongyoon Hwang @dongyoon_ai]

Introducing "3D HAMSTER", accepted to IROS 2026! 🎉

Hierarchical VLA planners draw waypoints in 2D, but robots act in 3D. Give the VLM a depth encoder, and it predicts metric 3D trajectories robots can execute.
Project page: https://t.co/5TiGDOmHLB
Paper: https://t.co/Rq3rIkUAP4

[Byungkun.Lee @Byungkun_Lee]

Introducing "See like a Robot"🤖
Robot data spans diverse camera viewpoints, making learning harder. Give a VLA robot-centric pointmaps, and it performs better with one extra encoder + one element-wise addition.
Project page: https://t.co/k5uBJOUBYa
Paper: https://t.co/ff2pwLLxxl

[Kangwook Lee @Kangwook_Lee]

Welcome to Seoul, ICMLers!

@Krafton_AI will be presenting many interesting papers throughout this week, so please check them out!

I will also help present some of the work (see 🧵below), so please come and say hi 👋

[Kangwook Lee @Kangwook_Lee]

https://t.co/nIvkmeIQFD