Where atoms
meet intelligence.
We build autonomous robots that work the field row by row, AI software for the industries that feed and move the world, and an ops team that keeps it all running while you sleep.
Something ancient is about to get an update.
The field doesn't know what's coming.
Neither does hunger.
What we build.
Hardware, software, and the people that run them — under one roof.
Field Robotics
Autonomous ground robots for sowing, weeding, fertilizer placement, and targeted pesticide spray. Per-plant precision, 24/7 uptime, no operator.
AI SaaS
Vertical AI platforms for agriculture, logistics, and retail forecourts. Yield prediction, fleet routing, store ops — on your data, in your stack.
Ops Support
AI-augmented back-office, finance, and customer ops as a service. A team of agents and humans that runs the boring, important work.
Why now.
Three forces converging that make this inevitable, not optional.
U.S. farm labor shortage hit a 20-year high in 2025. 60% of large-scale producers can't find workers. Crops are rotting in fields. The workforce isn't coming back — the average farmworker is aging out, and enforcement is accelerating.
EPA's final glyphosate registration review lands October 2026. State-level bans are already fragmenting the market. Herbicide-dependent farming is a liability. Mechanical and precision alternatives aren't optional anymore — they're the compliance path.
Ag robotics funding hit $744M in 2024. Autonomous weeding drew more precision ag venture capital than any other category in 2025. Ecorobotix closed $105M. Carbon Robotics raised $70M. The market has decided this is real.
Where we are.
Honest numbers. This is the build phase, not the pitch deck version.
One stack. Three surfaces.
The robot, the model, and the operator share the same brain.
Hardware that learns.
Software that moves.
Most companies sell you a robot or a dashboard. We sell the loop: robots in the field generate per-plant data, our models turn it into decisions, and our ops team makes sure both keep running. The same perception stack that guides the robot prices your harvest.
- Per-plant detections feed the same forecasting models that plan your season.
- Logistics routing reuses the same vision stack that navigates the rows.
- Ops agents triage exceptions — humans see only what matters.
“Racetrack-style forecourts are the highest-velocity retail in the country. 800+ stores running on instinct instead of data. Per-SKU, per-hour demand forecasting is the unlock nobody's shipped yet.”
“Mid-market carriers run 40-truck fleets with Excel and phone calls. They can't afford an OR team. A drop-in dispatcher that re-solves routing every 30 seconds against live HOS and weather — that's the gap.”
Who's building this.
One person with a thesis, the right background, and an unreasonable timeline.
Karthik Chowdary Tsaliki
Staff ML Engineer who's shipped production AI at scale — fraud detection systems that saved millions in chargebacks, payments security infrastructure at global scale, and AI-driven threat prevention using LLMs to build proactive security agents.
Now pointing that same stack at the field, the fleet, and the forecourt. Practices Vipassana. Believes the best technology disappears into the work it does. Based in Austin, TX.
Working on something at the edge
of farm, fleet, or store?
We're talking to design partners and early investors across all three lines. If you have soil, trucks, or a forecourt — and a problem that doesn't have a vendor yet — let's talk.