Projects

Selected work shaped by real operating conditions.

These examples show the kind of work Pixtell takes on: deployments and system design efforts where the technical solution has to fit the business environment, not the other way around.

Agriculture Drone imagery Pixtell product

FieldCoPilot — AI-powered field monitoring for farmers

A Pixtell product applying computer vision to drone imagery for continuous monitoring of meadows and crops. Designed to give farmers actionable visibility across their fields without manual inspection at every growth stage.

FieldCoPilot drone field monitoring imagery

Operating environment

Outdoor agricultural land with seasonal variation, changing light conditions, and diverse crop and meadow types across variable terrain.

System scope

Drone image processing pipeline, detection and classification logic, and structured outputs designed to integrate into farm management workflows.

Business relevance

Reduce the time between field observation and decision, and lower reliance on manual walk-throughs across large or fragmented land areas.

In collaboration with

Agriculture Edge deployment Jetson

Field weed detection for targeted treatment

Visual AI running on tractor-mounted NVIDIA Jetson hardware to detect weeds in motion and support localized treatment decisions under variable light, vibration, and field conditions.

Operating environment

Open-field deployment with shifting light, dust, plant overlap, and tight latency expectations.

System scope

Detection pipeline, hardware optimization, and reliable handoff from perception to machine action.

Business relevance

Support more precise intervention and reduce reliance on broad treatment strategies.

In collaboration with

Environment Drone imagery Field mission

Penguin monitoring in the Antarctic sea

Detection workflow for penguin counting from drone imagery in collaboration with Jean-Louis Etienne and Agroscope, designed to produce operationally usable outputs for field-linked monitoring activity.

Operating environment

Aerial imagery with challenging scale, variable backgrounds, and mission-specific reporting needs.

System scope

Detection logic, data structuring, and count-ready outputs that can be reviewed and used by non-technical stakeholders.

Business relevance

Improve the speed and consistency of wildlife monitoring workflows tied to environmental research programs.

In collaboration with

Enterprise Advisory Procurement

Requirements and acceptance design for enterprise roll-out

Visual AI requirements work that translates a business need into deployment-ready specifications, acceptance metrics, and vendor governance before implementation starts.

Requirements and acceptance design project imagery

Operating environment

Large-organization procurement where unclear requirements create delivery risk and vendor misalignment.

System scope

Use-case framing, technical criteria, integration expectations, deployment logic, and measurable acceptance thresholds.

Business relevance

Reduce implementation risk and give buyers clearer control over what must be delivered and how it should be evaluated.

In collaboration with

Across projects

What stays constant from one deployment to the next.

1The system is defined around a business action, not just a technical benchmark.
2Deployment conditions are treated as first-order design inputs.
3Integration decisions are made alongside model decisions, not after them.
4Clients need a practical path to run, review, and improve what gets delivered.