Most robots on a competition floor are built to do more: move faster, lift heavier, decide on their own. Tessalt was built to do something harder. It finds the match, proves the match, and then stops.
“Everyone expects a robot to finish the job,” Shiv Mandlik says. “Ours is designed not to. That’s the whole point.”
India’s heritage has a scatter problem. Centuries of looting, dismantling and piecemeal excavation have left single sculptures split across collections that have never compared notes. Reuniting them has depended on a curator’s memory and a great deal of luck.
Tessalt, which Mandlik co-built in 2026, takes the luck out. The system scans each fragment robotically, photographs it from angles it chooses adaptively instead of from a fixed rig, runs the images through calibrated computer vision and rebuilds the piece in 3D. Every proposed match then has to survive six stages of verification before anyone acts on it.
The sharpest idea in the system is the Fragment Passport, which Shiv Mandlik designed. Every piece carries a digital record of four things: geometry, provenance, permissions and candidate matches. A matching algorithm needs only the first. Mandlik insisted on the other three.
“Making two edges fit is the easy part,” he says. “Where has this piece been? Who gets to decide what happens to it? If you can’t answer that, a perfect match is still the wrong move.”
Then comes the decision that separates Tessalt from the rest of the field. The machine restores nothing. It proposes. The final call on every reconstruction stays with a human expert. Full automation would have looked better on stage, and a teenager with four granted patents could easily have sold it. Shiv Mandlik built the opposite.
“If the robot’s wrong, you’ve damaged something that survived a thousand years,” Mandlik says. “I’m not handing that call to an algorithm. The machine does the searching. The person who has spent a lifetime with these objects does the deciding.”
The approach was tested on real history, not a staged demo. Working with some of India’s most renowned temples and museums, alongside the Archaeological Survey of India and Gujarat University, Tessalt helped identify and reassemble the six artefacts. The WRO India bronze followed.
The instinct behind Tessalt did not start in a lab. It started at a potter’s wheel.
Mandlik’s family were potters once, then walked away from the craft. Seven generations passed before anyone went back. Shiv Mandlik did, apprenticing under master potters in Kutch while carrying a full IB Diploma and a construction-safety nonprofit.
“The first month, everything I threw collapsed,” he says. “Every single piece. That’s when it clicked. The craft never left my family. We just stopped listening to it.”
A potter learns early that clay keeps a record of every mistake, and that an object breaks along the way it was made. Read the break and you know where the missing piece should meet it. That is, in plain terms, the question Tessalt answers.
“I don’t look at a fragment and see a puzzle piece,” Mandlik says. “I see whoever made it. The wheel taught me that before any code did.”
He has not treated the craft as a private hobby. Sixty-plus pieces in, he has recorded oral histories from more than a dozen elders, pulled 40 community members back to the wheel and mentored 35 to 40 young people in it. The robotics work has also turned back toward its source, aimed at reintegrating rare pottery traditions.
Read together, it is one problem at two scales. A family broke from its craft. A civilisation’s sculptures broke from each other. Shiv Mandlik has spent two years putting both back together, one at the wheel and one in the lab, and has a national medal to show for the second.
At the wheel, a potter who forces the clay gets a crack. Mandlik has learned what plenty of engineers twice his age have not: build the tool, then know when to take your hands off it.
