10/ At ICPP 2026 in Singapore, a leading venue for parallel and distributed systems, our Chief AI Scientist @zehuac presented our paper on learned gradient sketches. Communication is a core bottleneck in decentralised training. Rather than sending full gradients between nodes, we learn what needs to be sent. Reply to @flock_io: 9/ Our research also reached a wider audience. Our case study on collaborative AI in glucose monitoring through federated learning was featured in @wef’s MINDS, a global programme on real-world AI applications. https://x.com/flock_io/status/2077264777076982198