Odin Robotics brings adaptable robot arms to existing factory lines

A robot that needs a factory rebuilt around it can be a hard sell to a smaller manufacturer. Odin Robotics is approaching the problem from the opposite direction: installing robot arms into existing lines to handle repetitive work as products and conditions change.

The startup’s launch focuses on machine tending, packing and palletizing. Its proposed customers are small and midsized factories whose workers move parts between machines or pack products by hand, often because their operations don’t suit a rigid automation setup.

Odin’s demonstrations make that pitch tangible. In one, a robot tends a CNC mill, loading raw parts, running the machining cycle and removing finished pieces. Another shows food containers being collected from a moving conveyor and placed into totes. The company presents these as demonstrations, without disclosing measured customer production results.

Watch Odin Robotics’ original launch video.

The technical approach combines programmed controls with learned skills. Fixed controls handle steps that require predictable execution, such as interacting with machine doors, fixtures and defined safety zones. Learned behavior handles variation in parts and their placement, including mixed products or irregular objects.

That division is central to the product. A factory may want a door opened the same way on every cycle while needing the robot to recognize a part that arrives at a slightly different angle. Odin says its systems are designed to cope with that variation rather than requiring every object to be positioned perfectly in advance.

The company supplies the complete installation: the robot, its end effector, sensors and integration with the existing line. Prospective customers show the team a task and its current volume; the parties then agree on scope, price, timing and safety requirements. The launch does not provide standard pricing or a promised installation timeline.

Odin argues that each deployment should contribute reusable skills and machine integrations, reducing the work required for the next factory. That is a proposed scaling advantage, rather than a quantified result in the announcement. The practical challenge is how much of one installation can carry over when the next customer’s equipment or product mix differs.

Co-founders Samriddhi Singh and Alper Sametoglu came to the problem through hands-on engineering. Singh, also known as Simi, founded GirlsCanCode at 13 and says it grew into a government-funded program. Her experience includes humanoid and industrial robots; she left electrical engineering studies at ETH Zurich to build Odin.

Sametoglu worked at a manufacturing-intelligence startup while in high school and later built rockets and satellite propulsion components. Their company is part of Y Combinator’s Fall 2026 batch.

The official website shows accelerated previews of its CNC and conveyor demonstrations, clearly labeled at three times and 2.5 times speed. Those clips illustrate the workflows, but they aren’t a substitute for cycle-time measurements in a customer’s plant.

For now, Odin is asking manufacturers to bring it specific repetitive tasks. That keeps the next step concrete: show the team the loading, packing or palletizing station, then establish whether a robot can fit the job and the line already in place.

‍