👥 Looking for Teammates? Reply Here!

If you’re new and feeling a bit lost, don’t fret! We recommend teaming up with fellow AIcrowd users for a more enriching experience!

Reply to this thread with a brief intro about you, what brings you to this challenge, and see the magic happen :sparkles:


:rainbow: Welcome thread

4 Likes

Hey, I’m looking for teammates, a bit about me

  • I’m Rajath Bharadwaj, a CS undergrad with good experience in DL and ML
  • Looking forward to learn and contribute
  • RL has been my fav domain, so NetHack will be exciting

I like playing around with different RL algorithms and making my own custom env. I want to apply what I learn here to my real-world projects. I want to colab with my fellows and have fun building this and also learn in the process. I guess that’s all I have to say :smile:

2 Likes

My name is Clint, and I am looking to team up with someone who is more familiar with RL than I am (but perhaps not as familiar with Nethack itself?).

The last time I did any Nethack programming was with nHackBot back in 2005 / 2006. At the time, the Nethack bot community spent a lot of our time fighting through the terminal interface, and so seeing what NLE provides is such a breath of fresh air. I feel like NLE does a great job of handling all of the difficult pieces, and lets bot coders really concentrate on the “fun stuff”. :slight_smile: Huge thanks to the NLE team for creating this – I’m extremely excited about it! As far as machine learning, I only have a little experience in using ML classifiers, but I haven’t yet done any RL and I’m very interested in learning how to apply it.

I have a fair bit of experience with Nethack – having done many regular ascensions, as well as several conducts (including multiple pacifist ascensions, and one blind / “zen” ascension that I’m particularly proud of).

I’m interested in leveraging my knowledge of Nethack to team up with someone who is more familiar with RL. I’m not entirely clear how it would work out, but one possibility is that I would perhaps be primarily responsible for providing filtered game state pieces or inferred data that my teammate would then feed into a sensible RL structure. I am looking to take an “informed” / “hybrid” approach to the challenge (where we will write some hand-tuned heuristics or best-practices into the bot) rather than a more “pure” ML approach.

Thank you!

1 Like

Hello, looking for the collaborators.

I am PhD student in Particle Physics, I have a lot of experience in Programming and ML.
Looking for the productive, fun collaboration, ready to share the knowledge and help others.