Kaelor Labs

Company

A laboratory for physical-world intelligence.

Kaelor Labs is the company behind MissionSmith. We build AI-native systems that help autonomous machines understand environments, reason through complex conditions, and execute real-world missions.

Thesis

Autonomy begins with understanding the environment.

Most AI still lives on screens. It writes, retrieves, and recommends. The physical world is a different problem: geometry, constraints, machines, and consequences that leave the interface.

Kaelor Labs is not a drone company. Drones are the first application. The longer work is the intelligence layer that lets autonomous systems perceive a place, reason through intent and limits, and act.

MissionSmith is the first commercial product. It is a wedge into professional drone operations — not the whole company.

People

Founders

  • Lourd Arun Raj

    Co-founder & CEO

    AI researcher focused on agentic systems, multimodal reasoning, and computer vision. Inventor on five granted patents, with experience taking advanced AI from research into production. Leads Kaelor Labs’ agent architecture and spatial intelligence, keeping the company on systems that understand environments, plan action, and stay inspectable before they are trusted.

  • Ankit Luthra

    Co-founder & CTO

    Has delivered AI systems and software services for enterprise clients, including Microsoft, building models and tooling that move from prototype into production workflows. Background across applied AI, product engineering, and systems you can inspect, not just prompt. Leads Kaelor Labs product and technical architecture: spatial intelligence, geometric verification, and the systems that let an agent propose a plan without being allowed to certify it.

Operating

How we decide what to build.

  1. Verifiability over vibes

    A plan is trusted when another engineer can inspect the checks, not when the model sounds confident.

  2. Honest limits

    We do not invent customers, certifications, or capabilities that are not shipped. Early work is described as early work.

  3. Physical-world rigor

    Geometry is unforgiving. Collision and clearance have to be computed against the actual environment, not a flat map.

  4. Human authority

    Nothing flies, drives, or actuates until a human approves. The operator in command remains responsible.

  5. Room to grow

    The company is built for more than one product. MissionSmith is first, not final.

Direction

We build the intelligence layer for autonomous systems.

Inspection, infrastructure, surveying, robotics, and other physical operations all need the same loop: world, understanding, reasoning, action, feedback. We will not pretend that loop is finished. It is the direction of the work.