Aryaman Reddi

PhD Student at TU Darmstadt

[Nico Bohlinger @NicoBohlinger]

⚡ One policy, millions of embodiments, over 200 robot models. Can we add yours?

We're building γ₀, a generalist RL policy for motion control trained across millions of randomized embodiments derived from a growing collection of more than 200 robot models.

[Steve McCormick @Quasilocal]

Couldn't help myself

[Aryaman Reddi @ReddiAryaman]

Our paper on Streaming RL under Partial Observability won Best Paper Award at the Big Worlds Workshop at #RLC2026. We enable memory in streaming RL using recurrent trace units.

Project page: https://t.co/gXGExiEIIz

Collaborators: @__noahfarr__ *, @CarloDeramo , @Jan_R_Peters

[Emil Ares @emilares]

$6.7T is going toward data center capex through 2030. It still takes 10-12 months to design data centers.

@MarengoYCS26 is the engineering firm for accelerated data center development. Site Due Diligence through FEED and Permitting Design in half the traditional time and cost.

Data center design takes 10-12 months, and because every concept is manually produced, design firms explore only 3-4 options per site, leading to suboptimal designs, risks that surface late, and delays that cost developers.

Marengo reduces those traditional pre-construction cycles down to 5-6 months by automating iterative engineering processes, allowing for faster design iterations, earlier constraint and risk visibility, and higher-confidence decisions, turning early development speed into a lasting advantage.

Our professional engineering team uses in-house-built AI tools to accelerate the work, validating and choosing the best designs for you. They present you with a decision-ready package built from data and expertise, at every stage: Due Diligence, Feasibility, Concept, and FEED/Permitting Design.

Before YC, @gadmarconi and I helped build and deploy on-site data center infrastructure, worked on the world's most powerful magnets (56MW) at the 1.4GW MagLab facility, spent time as an AI software engineer at a nuclear and energy EPC, as well as published multiple papers on AI for complex engineering design. We know firsthand the manual processes that slow down design and lead to suboptimal decisions.

Are you a data center developer? Send a site. Let us show you what accelerated engineering looks like on one of your projects. https://t.co/mPaIKV5KGX

[nowherelad @nowherelad1]

all bikinimorrows

[Aryaman Reddi @ReddiAryaman]

Scientists really called this guy 'sleepy small brain'

[Aryaman Reddi @ReddiAryaman]

This paper has interesting evidence about what may be the first instance of bipedalism in hominin evolutionary history:

https://t.co/al8dS5vHT9

[Aryaman Reddi @ReddiAryaman]

https://t.co/L1fKLPgHVY

If the shortened url doesn't work^

[Aryaman Reddi @ReddiAryaman]

This article on respiratory systems is really interesting. I had no idea that birds and crocodiles shared a breathing system that goes all the way back to the common ancestor of all archosaurs.

https://t.co/xBfUzui1Va

credit: @EBKaczmarek, J. Phillips, B. Ryerson

[Aryaman Reddi @ReddiAryaman]

The best shot from Jurassic Park.

[Aryaman Reddi @ReddiAryaman]

Very interesting research from my friend @RKrohn_! How can we use multiple sensors to improve the efficiency of online deep RL training?

[Rickmer Krohn @RKrohn_]

Real-world online Reinforcement Learning from Vision + ForceTorque + Proprioception --> With MSDP we achieve ~90% success rate in only 35min of online training!

Checkout the Preprint here: https://t.co/bmme0ESX0v

Thanks to @FatAndFurious42 @GabrieleTiboni1 @GeorgiaChal!

[Aryaman Reddi @ReddiAryaman]

Totally stoked that people like our paper! If you want to know more about recursive reasoning in multi-agent RL, come to my poster at #EWRL2025 at 11:00 ♻️

[Francesco Bertolotti @f14bertolotti]

In this paper, the authors compute the gradient update of the policy of one agent by accounting also for the update of all other agents. I feel this is a fairly general idea that could be applied to most multi-agent RL algorithms.

🔗https://t.co/nLLmmgpMIh

[Aryaman Reddi @ReddiAryaman]

@f14bertolotti Thanks for sharing our paper, I'm glad to see that people find the idea of recursive reasoning interesting!

I'll be presenting this at the European Workshop on RL this week for anyone who's interested in knowing more.

[Aryaman Reddi @ReddiAryaman]

@davide_tateo @liu_puze All the best Davide!

[Aryaman Reddi @ReddiAryaman]

What does BAMF mean?

In Germany it means 'Bundesamt für Migration und Flüchtlinge' (Federal Office for Migration and Refugees), but whenever I hear it, I can't help but imagine the sound Nightcrawler from X-Men makes when he teleports

#germany #xmen #comics

[Théo Vincent @Theo_Vincent_]

Very excited to present 🎉Eau De Q-Network🎉 on Thursday at RLDM Poster #28

🔍Eau De Q-Network gradually prunes the network weights at the agent's learning pace, ultimately reaching a final sparsity level that is discovered by the algorithm!🔎

👉📰 https://t.co/4MV6noQ7Hd

[Aryaman Reddi @ReddiAryaman]

"...for there is no greater glory that can befall a man living than what he achieves by speed of his feet or strength of his hands".

The Odyssey, book 8 line 147 (Richmond Lattimore's English translation)

#homer #odyssey

[An Thái Lê @an_thai_le]

Stuck in local minima? We introduce tensor sampling + splines for high-entropy, smooth control trajectories, enabling high-entropy sampling-based MPC 🚀

Preprint: https://t.co/aiucqHyHaV
Website: https://t.co/Ejk4lBuxuY
Code: https://t.co/5tVgmwBptY

#robotics #MPC #control

[Théo Vincent @Theo_Vincent_]

How do you tune the hyperparameters of your RL agent? 🤔

Come and chat with me about it tomorrow afternoon at poster 397 of @iclr_conf !

I will be presenting⚡️ Adaptive Q-Network⚡️a method that adaptively selects the hyperparameters of your RL agent during training 🔥