From a final-year project to factory floors
Four batchmates, one cloth-folding robot, a Singapore apartment, and a decision to move to India. The story of how Piqoid started — and why we believe the timing is finally right.
We met as batchmates in robotics at NTU Singapore. For our final-year project, under Prof. Pham Quang Cuong of Eureka Robotics, we built a robot that folds cloth — one of the classic hard problems in manipulation, cloth being deformable, slippery and endlessly non-repeatable.
The project taught us two things. First, that the hard problems in manipulation were tractable with the right stack. Second, and more important: the hardware was already affordable. The arms weren't the bottleneck anymore. What was missing was the intelligence to deploy them outside carefully engineered labs.
Nights and weekends
We started building nights-and-weekends in a Singapore apartment. Aditya had led a 12-person team at Unilever (€20M+ in savings) and shipped AI agents at TCS. Jainil had built an engine-inspection robot for Singapore Airlines and the world's first 3-in-1 cleaning–security robot — and grew up in a family-run Indian manufacturing SME, which meant he knew exactly who these robots were for. Ahan brought software, AI and firmware (MS Robotics, Johns Hopkins). Mohil, an ex-plant manager from the Emami Group, kept us honest about what a real plant will and won't accept.
The pull from India
The surprising part wasn't building the robots — it was who kept calling. Indian factory owners, family friends of Jainil's family, small fabrication shops. They weren't asking for futuristic machines. They were asking: can a robot load our press brake? Can it weld this bracket? Because we cannot find people to do it.
Unskilled labour runs 10–20% unannounced absence on any given day. Skilled welders and cutters are scarce and retiring. The next generation doesn't want to run manual factories. And the automation industry — the system integrators — is structurally incapable of serving these factories at their price point: months of quotes, 3–4× the robot's cost in engineering.
We felt a strong pull, and in June 2024 we moved to India full-time. Ahmedabad first: the press-bending cluster, the machine makers, the factory visits that taught us that ~40% of machines have no controller at all.
Where we are now
Today our stack runs end to end: the arm, the modules, the control suite, and on-robot VLA policies with dynamic replanning. We've demonstrated it on real factory tasks — press-bend tending, unstructured 3D metal cutting and welding, voice-visual sorting — and we're interviewing IIT and NIT engineers to grow the team.
In 15 years, every factory will own robots. We intend to make that happen in India — one task, one city, one compounding deployment at a time.
— Aditya & Jainil · work with us →