We're a small team building the science of organic alignment—the claim that the kinds of agents that stay aligned with us are the kinds that stay aligned with each other. We come from backgrounds in AI safety, developmental biology, infrastructure, and collective intelligence research, brought together by our mission.

Founders

Emmett Shear

CEO & Founder

• Partner at YC, founder and CEO of Twitch, and extremely interim CEO at OpenAI
• Independent researcher in alignment and agency before cofounding Softmax

David Bloomin

Founder

• Built large-scale infrastructure at Google, Facebook, and Asana for two decades
• Cofounded Plurality Institute and led research on multi-agent AI frameworks and governance methods

Board

Adam Goldstein

Board Member and Founder Emeritus

• Cofounded Hipmunk before joining Michael Levin's lab at Tufts to research cell signaling, bioelectricity, and basal cognition
• Cofounded Softmax and brought the organic alignment thesis into the company's multi-agent research

David Langer

Board Member

• Founder and Managing Partner at Lionheart Ventures, backing AI, mental health, and climate companies
• Previously founded Zesty (YC-backed, acquired by Square in 2018) and GroupSpaces

Team

Yatharth Agarwal

Cofounder Stanford AI Group · Screentime addiction researcher

• Created first high-school curriculum and for-credit Stanford class on AI safety; co-taught AI ethics course attended by Duke/NYU faculty; led AI regulation seminar at Stanford Law with CA Supreme Court justice
• Previously led AI at Axle Health; researched screentime addiction; and written extensively on self-alignment in humans

Malcolm Handley

Ex-Asana #1 · Founder, nuclear fusion VC

• First employee at Asana, early engineer at Google
• Founded Strong Atomics, a nuclear fusion VC; advisor at ARPA-E

Kyle Herndon

ML compilers · IREE, torch-mlir

• Software engineer specialising in ML compilers and high-performance systems, with contributions across IREE, torch-mlir, and the ROCm ecosystem
• Works on performance engineering and the multi-agent reinforcement learning training stack

Richard Higgins

PhD Michigan · CVPR, ICCV, NeurIPS

• PhD at Michigan, with research spanning compositional image editing with latent diffusion (ICCV 2025), hand-object segmentation in video (CVPR 2023, NeurIPS 2021), and solar magnetic field estimation
• Previously visiting academic at NYU Courant and CV research intern at Meta

Malcolm Ocean

Writer · Creator of Intend.do

• Writer, software designer, and researcher focused on collective intelligence, rationality, and the mechanics of trust between agents
• Created Intend.do, a goal-pursuit app rooted in his writing on meta-cognition, and has spent over a decade publishing influential essays on mutualism, cultural evolution, and how groups think together

Subhojeet Pramanik

Alberta RLAI · ex-IBM Research

• MSc at Alberta's RLAI lab; AI researcher in reinforcement learning and representation learning; introduced the Recurrent Linear Transformer (AGaLiTe, TMLR 2024)
• Previously at IBM Research and author of OmniNet, a unified architecture for multi-modal multi-task learning

S. Alex Smith

PhD UCLA · Led infra at Asana

• Math at Cornell, then math PhD at UCLA
• Led the infrastructure organization at Asana with deep experience in large-scale backend and storage systems

André von Houck

Creator of pixie, boxy, chroma

• Prolific contributor to the Nim programming language community, where he has authored many widely used open-source libraries — pixie (2D graphics), boxy (GPU rendering), jsony, chroma, and 40+ others
• Works on the game engine and game UX