OUR STORY
Built from a practical question: how much sensor data do you actually need?
Lightscline began by helping legacy machines become intelligent. The storage problem we encountered became the starting point for a much larger mission: make high-frequency sensor intelligence fast, efficient, and deployable anywhere.
Today, Lightscline is building the next-generation on-site AI infrastructure layer for industries where sensing is expanding rapidly, decisions are time-sensitive, and compute, power, storage, transmission, and latency are constrained.
CO-FOUNDERS
Built by experts across industrial AI, sensing, and large-scale systems.
FROM MACHINE RETROFITS TO SENSOR AI
A constraint became the company.
In 2019, Ankur Verma and Ayush Goyal began working on a project to retrofit legacy industrial machines with sensors. Their goal was to create an open platform that small and midsize manufacturers in India could use to monitor equipment and predict failures.
But the project exposed a fundamental problem: collecting dense sensor data was easy; storing, moving, and interpreting all of it was prohibitively expensive for the very businesses the platform was meant to serve.
That insight changed the direction of the work. Instead of asking how to build more infrastructure around an ever-growing stream, the team began asking which fraction of the data actually carried the information needed for a decision.
THE LIGHTSCLINE JOURNEY
The first constraint
Work on sensor-retrofitted legacy machinery reveals that conventional data collection and storage are too costly for practical deployment.
A new technical thesis
Ankur, Ayush, and Penn State industrial engineering professor Soundar Kumara bring together sensing, predictive analytics, and large-scale software design to identify the most informative sensor data.
From hypothesis to platform
NSF I-Corps, Penn State LaunchBox, the Diefenderfer Fellowship, peer-reviewed research, and customer deployments pressure-test both the technology and its market need.
One layer across industries
Lightscline now supports maritime, manufacturing, and energy workflows—turning terabytes of multi-modal sensing into near-real-time intelligence from edge to cloud.
HOW WE WORK
Independent thinking. Rigorous testing. Rapid iteration.
Start with the physics
Understand the instrument, environment, and decision before choosing a model.
Prove it on real data
Benchmark openly, validate across deployments, and let evidence—not scale for its own sake—guide the architecture.
Design for deployment
A useful model must fit where decisions happen: on site, on the edge, on premises, or in the cloud.
OUR MISSION
Make advanced sensor intelligence economical and accessible—without requiring every organization to build hyperscale infrastructure.
BUILD WITH LIGHTSCLINE