Simular builds a local AI agent designed to run on personal devices and automate routine work tasks. The stack—Python, PyTorch, JAX, React, SwiftUI—reflects a machine-learning foundation paired with native desktop and mobile frontends. Active projects span deployment challenges (real-world agent deployment, VM orchestration), multimodal reasoning, and a backend cloud service, signaling early-stage friction between local-first architecture and cloud infrastructure needs.
Notable leadership hires: Community Lead
Simular is a two-year-old AI agent startup based in Palo Alto, founded by Ang Li and Jiachen Yang. The product runs on user devices to automate manual workplace tasks—email handling, task management, scheduling—without offloading computation to the cloud. The team is small (2–10 headcount) but distributed across engineering, research, marketing, and design roles, with active hiring in the United States and Singapore. Pain points include real-world deployment reliability, customer-specific adaptation, and scaling backend services.
Python, PyTorch, JAX for the ML core; JavaScript/TypeScript, React, and SwiftUI for frontend; Go and Rust for infrastructure. The stack emphasizes local execution and native platform integration.
Core projects include real-world agent deployment, multimodal grounding for reasoning, a cloud backend service, go-to-market strategy, and planning/RL systems for computer use automation.
Other companies in the same industry, closest in size