About
Kilnara exists to make robotic factories adapt as quickly as the products they build — safely, measurably, and without weeks of manual re-tuning.
Every robotic line should learn from yesterday and be ready for tomorrow.
We are building the closed loop between the physical factory and its simulation twin: observe the line, generate the hard cases, retrain the policy, validate safety, deploy at the edge, and feed telemetry back into the next run.
Our goal is simple: help manufacturers move from brittle automation to resilient Physical AI that compounds over time.
We built Kilnara after watching brilliant automation teams lose weeks to tiny changes.
Fixtures shifted, lighting changed, new SKUs arrived, and otherwise capable robots started missing picks. The missing layer was not another dashboard — it was a living twin that could create training data, test policies, and deploy only what passed.
Factories change daily
Modern lines face short runs, model-year churn, and SKU volatility that static automation cannot absorb.
Real failures are expensive
Collecting edge cases on live cells costs throughput and can put people or equipment at risk.
Simulation should operate
A twin should not stop at visualization. It should generate, train, validate, and improve the line.
From research loop to plant-floor platform.
Kilnara is young by design and disciplined by the needs of enterprise robotics teams.
Prototype twin loop
Connected line telemetry to synthetic scene generation and policy retraining in a lab cell.
Design-partner cells
Validated overnight retraining workflows with automation teams in automotive and logistics.
Safety gate hardened
Added threshold-based validation, approval routing, release records, and rollback.
Edge deployment
Packaged policies for low-latency inference close to robots, cameras, PLCs, and MES.
The operating system for our team.
We work at the boundary of software, robotics, and production operations. These principles keep the work grounded.
Reality first
Every model, twin, and metric must reconcile with the physical line.
Safety is a product
Validation, approvals, and rollback are core features, not implementation details.
Earn trust with evidence
We prefer measured outcomes, audit trails, and clear thresholds over vague claims.
Respect the floor
Operators and controls engineers know the line; our software should make them faster.
Compound learning
Telemetry should make the next twin, dataset, and policy better than the last.
Build for enterprise
Security, deployment flexibility, and maintainability matter from day one.
Builders from robotics, simulation, controls, and enterprise software.
Supported by builders of industrial AI.
Kilnara is backed by early-stage investors, robotics operators, and technical angels across manufacturing, simulation, and edge AI. We are pre-seed/seed stage and focused on design-partner outcomes.
Product decisions start at the cell and end with measurable plant impact.
We pair closely with automation teams, validate in simulation, and ship in controlled steps.
- Observe before optimizing — baseline the real process and failure modes first.
- Ship with rollback — every deployment has a known safe previous state.
- Design for operators — explain outcomes in terms of throughput, uptime, and risk.
- Keep data under control — support VPC, on-prem, and air-gapped environments.
Early signals from the robotics community.
Selected for technical showcases and cited in conversations about Physical AI, digital twins, and safe robotics deployment.
Help build the self-improving factory.
We are hiring people who can bridge rigorous robotics research and the constraints of production operations.
Open roles across robotics, simulation, and field engineering.
Join a small team where your work can reach physical production lines quickly and safely.
View careersBuild the next generation of adaptive automation with us.
Talk to Kilnara about design partnerships, careers, or bringing the loop to your line.