Company

About Yuji Labs

Yuji Labs Private Limited is an Indian deep-tech company. We build Yunify™, a physics-driven AI operating intelligence layer for energy plants, and we build it from inside the industry rather than adjacent to it.

Bhavik Modi

Co-Founder and Chief Executive Officer

Strategy and business development · 12 years

Twelve years in instrumentation and process engineering, electrolyser technology and machine learning, with infrastructure project management across multi-billion-dollar energy projects. Previously at Siemens, L&T, Mitsubishi and Newtrace, spanning industrial scale-up and large-scale clean energy deployments. He leads customer strategy, commercialisation and industry partnerships.

Deeparnak Bhowmick

Co-Founder and Chief Technology Officer

Technology and product · 15 years

Fifteen years in advanced process control, industrial automation, multi-physics modelling and safety systems, and the research and development behind the company's patent-pending work. Previously at GE, Mitsubishi and Alstom. Certified Automation Professional and Functional Safety Expert. He architects the Yunify™ intelligence engine and leads deep-tech product development and intellectual property.

Certified Automation Professional (ISA CAP)Functional Safety Expert (FSE)

Why we started this

We spent our careers inside operating industrial plants: power, oil and gas, renewables and most of what sits between them. We worked together at Mitsubishi, and again taking an electrolyser from laboratory to industrial deployment. Across all of it we kept hitting the same wall.

The control system raises an alarm. Hours or days later somebody works out why, by which point the efficiency and part of the asset life are already gone. Hundred-million-dollar assets are run on alarms, spreadsheets and phone calls.

Then every automation vendor started saying AI would fix it. It did not, and the reason was consistent. A model trained on history has no account of what the plant is physically doing, so it cannot separate a mechanism from a coincidence, and an operator stops reading it after the third false alarm.

So we left, and started Yuji Labs to build the thing we had wanted on site: physics first, AI on top of it, and an instruction at the end rather than another chart. It is not a demonstration. Yunify is deployed on operating plants across multiple continents, and it is paid for.

What we build

The variables that decide whether a plant earns money are largely not instrumented. A control system reports what a sensor measures, and its alarms are set to give an operator time to react to a condition rather than to describe how that condition developed. The largest recurring cost in most of these plants sits in something that degrades: an electrolyser stack, a battery module, a compressor train. Measured degradation data from operating plants is scarce enough that OEM lifetime claims go largely untested.

Capital cost gets the attention. Green hydrogen at around seven dollars a kilogram against roughly two for grey is the example everyone quotes, and almost all the effort to close that gap is aimed at capex. Operating cost is the least-addressed part of it. That is where the avoidable loss sits, in hydrogen and in every other asset we work on.

Yunify™ is our answer to that. It runs alongside the control system a plant already has, combines first-principles physics with AI anomaly detection, and produces a prescribed action. The intellectual property is developed and held in India, with a patent application pending and trademarks registered.

Where we work

One architecture covers the renewable value chain and the conventional plant beside it: green hydrogen electrolysers, battery storage, solar, wind, conventional gas power, and downstream compression, storage and ammonia synthesis under a variable hydrogen supply. It runs from a single megawatt to a gigawatt.

Our work spans green hydrogen, concentrated solar, onshore wind and conventional gas power, on deployments across multiple continents. One of them is an MW-scale green hydrogen and wind plant where Yunify is commercially deployed and running, and gigawatt scale is next.

The cross-asset coverage is deliberate. A hybrid plant with generation, storage, conversion and downstream equipment usually needs several monitoring vendors, none of which sees the interactions between them.

How we work

No modification to the installed control system. Yunify reads from the PLC, DCS or SCADA that is already there, over OPC UA, Modbus, MQTT or PROFINET, and does not alter control logic. The addressable base is every plant already operating, not only new build.

On-premise by default, though not only. Everything can run on a single industrial host on site with no outbound connection, which for assets with public or sovereign participation is usually a precondition rather than a preference. Cloud and hybrid deployments are available where a connection is wanted, and the plant-critical path stays local either way.

Vendor-agnostic, with no dependence on a single OEM. It is also why we take control systems and automation engineering work that has nothing to do with the product: the assumptions Yunify rests on get tested on other people's plants rather than on a roadmap.

These are constraints rather than features, and they were settled before the first line of Yunify was written. A product that assumes a greenfield plant, an outbound connection or a single OEM is a product for a plant that does not exist yet.

Frequently asked questions

Where is Yuji Labs based?

The registered office is in Surat, Gujarat, India.

Does Yunify work on a plant that is already operating?

Yes. It is designed as a retrofit. It reads from the existing PLC, DCS or SCADA over standard industrial protocols and does not require changes to control logic, so it does not depend on the plant being new build.

Which asset classes does Yunify cover?

Green hydrogen electrolysers in alkaline, PEM and AEM, battery storage, solar, wind, conventional gas power, and downstream compression, storage and ammonia. One architecture across all of them.

Is the technology developed in India?

Yes. The intellectual property is developed and held in India, with a patent application pending.