Research that has to work outside the lab
For me, building a company is where research has to face the real world. I have co-founded three ventures, and before any of them I built NAAMII from scratch as a profit-not-distributing institute, which now has 52 full-time staff.
Tangible
Tangible is building the infrastructure for reorganizing work in the post-AI world. An organization is the set of processes it runs, and a step in a process can now be taken by a person or by an AI agent. Tangible models what each actor can do, and uses that model for personalized, problem-based learning and for helping organizations redesign how their work gets done.
My part is the science: I lead the research, supervise the AI researchers, and help our engineers work in an AI-native way. Tangible was incubated within NAAMII.
- Personalized, problem-based learning
- Competency-based skills assessment
- Processes rehearsed alongside AI agents before they go live
- Research published openly, starting with The AI Pyramid (2026)
NAAMII
A research institute is a startup too. We structured NAAMII as a profit-not-distributing company so that it could earn, reinvest and outlive any single grant. Research becomes IP and talent, which then become either a spin-out or an R&D service for industry. It has grown to 52 full-time staff, with three deep-tech startups incubated to date.
Diyo.ai
Conversational AI and speech technology for Nepali and other under-served languages, built on NAAMII's language research. My role is an enabler in the ecosystem.
Terahs
The engineering and infrastructure side of health imaging AI. It works alongside my research on AI for ultrasound and medical imaging.
The thread across all of them: talent in Nepal is not the constraint; structure is. I talked about this at the National Young Academy of Nepal in 2022, under the title "Incubating start-ups in (non-profit) research institutes in Nepal". The NAAMII Innovation Factory is the next step: an attempt to make that structure repeatable. Read about it →