Deep Learning
Architectures from CNNs to transformers and state-space models, implemented from scratch, then scaled.
Kathmandu · Nepal · Worldwide
A community of researchers, engineers and students pushing deep learning, mathematics and AI research forward, from the heart of the Himalayas.
θ* = argminθ 𝔼x∼p(x)[ L(f(x, θ)) ]// about us
Frontier ML Nepal began as a small reading circle and grew into one of Nepal's most active machine-learning communities. We believe world-class AI research doesn't need a world-famous zip code. It needs curiosity, rigor and people who show up.
We go deep rather than wide. From the mathematics of optimization to the engineering of large-scale training, our sessions are hands-on, paper-driven and unapologetically technical.
// what we do
Architectures from CNNs to transformers and state-space models, implemented from scratch, then scaled.
Linear algebra, probability and information theory, the language behind every model, taught without hand-waving.
From SGD to second-order methods and learning-rate alchemy, how models actually learn.
Weekly paper reading, replication studies and original work, aimed at venues, not vanity.
// events
No events scheduled right now. Check back soon.
Follow our Facebook page for announcements and registration links.
// team
// join us
Whether you're writing your first line of PyTorch or your fourth paper, there's a place for you here.