TL;DR -
Twolabs builds humanoid robots for caregiving.
Our robots will assist at nursing homes and senior living centers to take care of the elderly, starting with feeding, dressing, medication, daily tasks, and companionship.
We have our first robot design Tobi, are actively training VLAs, and building the social intelligence layer. Check out our launch video.
What we do
Twolabs builds semi-humanoid robots for nursing homes and senior living communities. We're building the robot to achieve the following goals for the elderly:
We build the full stack in-house, the robot hardware and all the AI that makes it socially capable. Most of robotics right now is chasing warehouse manipulation, moving boxes and folding laundry. We think the harder and more important problem is robots and people working closely together.
The backstory
I'm Sardor. I was born in Samarkand, Uzbekistan, moved to the US at six, and studied CS at Cornell. At Cornell I led a 40+ engineer team building autonomous combat robots. We were ranked top 5% nationally; most of my team went on to defense tech and semiconductors. My faculty advisor was Professor Tapo Bhattacharjee, who runs one of the leading caregiving robotics labs in the country. After Cornell I joined Meta Superintelligence Labs at 23 years old, working on GPU optimization, foundation model training, and AI infra. I left to start Twolabs. (I also got to be the first Uzbek founder to get into YC.)
Danyal Ahmad is my co-founder, and we have known each other for over nine years, since high school. We interned together at Salesforce in 2024 and have built a lot of products together since. Danyal has an MS in AI from Georgia Tech, and fine-tuned video models at Salesforce.
We started Twolabs because our grandparents face a lot of health conditions as they grow older, and we wanted to build a robot that can assist them in their daily lives. The need is everywhere: the US is short hundreds of thousands of caregivers and the gap widens every year as the population ages. A caregiver costs a facility around $60K/year, turnover is brutal, and the work is physically and emotionally exhausting. Meanwhile the people who need help, getting out of bed, eating, not being alone all day, don't get enough of it. A robot doesn't replace caregivers. It takes the repetitive physical load off them so one caregiver can do more for more residents, and residents get help when they actually need it.
Our bet
Most robotics companies fall into one of two camps:
Each succeeds at one half of the problem. But we think a truly helpful robot has to handle both, completing the physical task and engaging with the person in front of it. Too many humanoids are optimizing for task completion alone. We're betting the robots that actually help people will need to be socially intelligent to do it.
The other half of our bet is how robots are built. Companies have to be vertically integrated to win in robotics right now. We build the hardware and the AI together, because at this stage of Physical AI the models need to be specific, not general-purpose.
Where we are
Prototype design completed and our AI stack in active development. We're advised by Professor Zackory Erickson at Carnegie Mellon, whose research is at the leading edge of assistive robotics.
Our asks
Contact us directly at [email protected] and [email protected]. And check out our website at twolabs.ai.