FARIS / ALLAFI

FOUNDER · HAMILTONIAN RESEARCH · ABU DHABI

Building what
comes after
attention.

Faris Allafi speaking at Make It In The Emirates in Abu Dhabi
FARIS ALLAFI · MAKE IT IN THE EMIRATES · ABU DHABI

The transformer is an important architecture. It is not the final one.

Hamiltonian Research studies models that replace quadratic attention and token-by-token generation with state-space dynamics and parallel refinement. The aim is practical: stronger computational scaling, bounded inference state, and more control over the tradeoff between generation time and quality.

DIMBA

Diffusion on a Mamba backbone.

A language model that refines a response in parallel instead of predicting it one token at a time.

DIMBA combines masked diffusion with a bidirectional Mamba-2 backbone. At each step, the model reads the incomplete sequence, predicts the missing tokens, commits the most confident ones, and repeats until the response is complete.

01 / INPUTPrompt + masked response
02 / DYNAMICSBidirectional Mamba-2
03 / PREDICTTokens + confidence
Commit the most confident tokensRepeat with fewer masks
Simplified DIMBA generation loop. The full implementation and experiments are public.

Faris Allafi

I’m an independent researcher and the founder of Hamiltonian Research, based in Abu Dhabi. My work is focused on efficient sequence models, non-autoregressive generation, and the systems work required to test those ideas properly.

I build the research and implementation together. Current work spans model architecture, training, inference, and low-level performance, primarily in Python, PyTorch, and Rust.

Connections

Research, systems, and difficult problems. Email is the most direct route.

DIP-8 CONTACT MAP · PIN 7 IS THE PREFERRED INPUT