Nantas Nardelli

Research, engineering, and other hats @ Carbon Re. Working on using and developing AI to mitigate climate change.
Previously found @ Oxford, DeepMind, FAIR, MSR, FiveAI, Edinburgh, [...]
🌍 / 🇪🇺 / 🇮🇹 / he/him.
Interested in working on AI + climate? Hit me up, I'm hiring!

Nantas Nardelli boosted:
Teemu RoosTeemuRoos
2022-11-30

I was a bit disappointed that there was no chance to ask questions at the invited talk by Alondra Nelson on the AI Bill of Rights.

I would have liked to hear how the White House expects that a document that (quoting as far as I recall) is not a "binding legal document" will have impact that is in any way comparable to the EU's corresponding regulatory actions (GDPR, AI Act, DSA, DMA, ...) which entail multimillion euro penalties when violated.

Nantas Nardelli boosted:
2022-11-18

I'm new here #Introduction

🔬 I'm a researcher at FiveAI focussing on uncertainty and robustness

🎓 Previously I was a phd student at Oxford Uni under Prof Phil Torr and N. Sid where I worked on multimodal and semi-supervised learning in VAEs

🏫 Cofounder of GirlsWhoML, an access program designed to improve gender diversity in #ai and #machinelearning

Please feel free to get in touch if you have questions or just want a chat

Nantas Nardellinantas@sigmoid.social
2022-11-13

Didn’t get to do any #rowing today, but London does look amazing on a foggy morning 🌁

Nantas Nardelli boosted:
Fabian Pedregosafabianp@mathstodon.xyz
2022-11-09

I'd like to migrate to Mastodon the Twitter bots I maintain such as @JmlrOrg and @TmlrSub.

What are good libraries / tutorials / examples for writing a (very simple) Mastodon bot in Python?

Nantas Nardelli boosted:
2022-11-08

Hello world. I'm a research scientist at apple / incoming prof at NYU interested in scaling up multi-agent learning to get it out there into the world, often in the context of transportation systems or autonomous vehicles

Nantas Nardellinantas@sigmoid.social
2022-11-08

@riedelcastro What a save!

Of course you can 🥰

Nantas Nardelli boosted:
2022-11-08

#introduction

Hi, I'm a research scientist at DeepMind studying structured representations and model-based deep RL. Previously, I did my PhD in Psychology at UC Berkeley in which I studied how people use mental simulation. The last few years I've been studying model-based systems like MuZero and trying to understand when they are better than model-free, and what role the model plays. I'm also very interested in structured reprs like programs and graphs and how to incorporate them into RL!

Nantas Nardellinantas@sigmoid.social
2022-11-08

@pfau maybe it’s time to force some kind of centralised community wide discourse…

Nantas Nardellinantas@sigmoid.social
2022-11-08

@pfau It’s kinda hard to tell whether Twitter has been objectively a net positive value add for us, but it probably depends on subjective experience. Plenty of people have used it to become community pillars (with both good and bad effects), and many junior members have occasionally managed to break through due to network effect. On the other end, it has also likely excluded folks who like to work outside mainstream ideas (but not all!). ¯\_(ツ)_/¯

Nantas Nardellinantas@sigmoid.social
2022-11-08

@pfau to me the big question is whether the AI community can fundamentally work as-is without a platform like Twitter. And if not, either (1) what do we need to do for Mastodon / Sigmoid to improve on the experience and maximise positive impact; or (2) how do we make AI research function at scale without a virtual public forum.

Nantas Nardellinantas@sigmoid.social
2022-11-08

@sharky6000 Don't think so, but the poll looks alright :)

Nantas Nardelli boosted:
2022-11-08

Alright (and more), curious to try a poll. What is your current position? (TIL polls capped at 4 options)

Nantas Nardelli boosted:
Pablo Samuel Castropsc@sigmoid.social
2022-11-07

hey everyone! this wednesday i'm going to be interviewed by sara hooker in a fireside chat organized by Cohere For AI.

really looking forward to this, and hope you can join us!

register here: zoom.us/webinar/register/11166

Nantas Nardelli boosted:
Nantas Nardellinantas@sigmoid.social
2022-11-07

I’m thrilled to announce that @carbon_re has raised a $4.8m seed round! We’ve been sitting on this news for a while to sort out all the legal bits, but we can finally share it around 🚀

It will help us grow the team, serve more customers, and kickstart an AI research group! However this is likely the first time you have heard of us, so let me tell you what we do.

Carbon Re is a #climate #AI company on a mission to decarbonise “foundation industries”. Why is this important, you might ask?

1/🧵

Nantas Nardellinantas@sigmoid.social
2022-11-07

@Ryder @sherif @chiara Great question!

Long story short, we are thinking hard about it, and it's very possible that we'll eventually do it, however it feels like the space is a little crowded right now (perhaps not in a good way, I might add...).

If you have any relevant contacts, we'd love to chat with people in the space to understand how best to start and quantify (+ maximise) the impact there!

Nantas Nardellinantas@sigmoid.social
2022-11-07

@sherif @chiara Adding @michael to the list, since he just joined us on Sigmoid!

Folks, time to get your #Introduction done!

Nantas Nardellinantas@sigmoid.social
2022-11-07

PS: @sherif and @chiara, some of my amazing coworkers, are also here on Sigmoid, so please feel free to go holler at them as well ❤️

Nantas Nardellinantas@sigmoid.social
2022-11-07

WE ARE #HIRING! 🎉

We need #ML engineers and scientists, #software engineers, physicists, chemists, #materials scientists, product people, and many more folks who really want to make a strong, positive impact against #climatechange.

We’ll soon be putting out jobs at apply.workable.com/carbonre, but in the meantime PLEASE HIT ME UP!

DM me or drop a quick message at nantas@carbonre.com, and I’ll be very happy to chat about how you could maximise your impact in #climate by working with us.

7/end

Nantas Nardellinantas@sigmoid.social
2022-11-07

Carbon Re is developing tools that use our collective knowledge about physics, chemistry, and machine learning to improve the efficiency of these industries such that their #GHG emissions can quickly decrease.

Straightforward, but not easy 😃 And this is why, friends, we need your help.

This data is largely new. Very few people have modelled these processes, and fewer have thought about automating them from a control perspective. Lots of hard, exciting new problems await us!

So… 6/🧵

Nantas Nardellinantas@sigmoid.social
2022-11-07

So that’s how @carbon_re comes into the picture.

We develop #AI based tools that enable these plants to improve their operations, aiming to get to optimal control policies. Cement produced in an optimally controlled plant generates significantly less CO2 than an average human-controlled plant, and there are only ~3000 of them to fix up!

It is more #sustainable, safer for workers, and pushes the industry to improve across the board. Improvements are good, and we need more of them, faster!

5/🧵

Map of 3000 cement plants in the world.

Client Info

Server: https://mastodon.social
Version: 2025.04
Repository: https://github.com/cyevgeniy/lmst