Lossfunk

10 projects of our batch 5 residency covered by Caleb Friesen in his latest video 🚀👇

Kickstarted Batch 5 of our six week AI residency. Next batch starts September 2025. Applications open!

Yep, we're building a third space for AI beyond academia and industry. We're an experiment to discover what could organizations of the future look like!

Update on Lossfunk research internship. 730 applications -> 150 shortlists -> kickstarted 10 projects already! Kinds of projects: • LLMs in prediction markets • World models for Atari-100k • Soft rewards (perplexity based) v/s Hard rewards • Meta-RL (learning to learn)

Our notes on basics of Reinforcement Learning! Crossposted on our substack: Link in Comments Please subscribe!

Hello Everyone,   Be part of an insightful talk this Friday, 20th June 2025.   Topic: Why Deep Learning Works So Well (even with 100 data points!) by Paras Chopra   Deep Neural Networks break the classic ML theory of bias-variance tradeoff. These models often have millions or billions of parameters, and yet they don't overfit the training data (as seen by continual decrease in test loss). What's happening? In this talk, we will build intuition about how deep learning works from a generalization point of view. We will talk about topics such as double descent, benign overfitting, flat basins, and so on. We will also fit a 1.8 million model on 100 data points, and show how it doesn't overfit!   (please request an invite first via the form below as we have limited space) Link in Comments

Hey Everyone, Be part of an insightful talk this Wednesday, 18th June 2025. Topic: Your Speech Model Isn't Bad - It's Just Built for Someone Else About the talk: Speech recognition often fails because it doesn't consider the diversity of real voices that arises from differences in age, gender, dialect, language and many other factors. Your own voice changes as you grow, or sometimes even when you change countries! In this talk, you'll discover why having a deep understanding of your end-user and building accordingly is crucial. We will explore this through my personal experience of building speech-to-text models in Gujarati and Hindi for children (primarily rural) in Gujarat and Rajasthan respectively. (please request an invite first via the form below as we have limited space) Link in Comments About the speaker: Ayush Deva  Current: ML Engineer @ Google Previously: ML Engineer @ Wadhwani AI, B.Tech - IIIT Hyderabad Date and time: 18th June, 6pm-7pm

So we finished the LeRobots Robotics Hackathon by Hugging Face over the weekend at Lossfunk. A lot of curious teams created amazing projects. Go check them out. Link in comments.

3D Printed Wall-e taking his baby steps at Lossfunk! We hosted Le Robot hackathon by Hugging Face!

A few hours into the Hackathon, It's gonna get crazier. Hugging Face