The power of private data sets

How does Privacy work for you?

AI model development is contingent on the available data. Accessing high quality data is a major challenge for many AI companies serving the agricultural sector as major concerns surrounding privacy and security prevail across the board. Consequently, building reliable and effective models is hampered as companies must invest excessive time and resources in their development and deployment.

TensoAI ensures secure and scalable AI model development: we are tackling the issues surrounding privacy, head-on, with a multitude of privacy-preserving techniques and end-to-end secure connection networks. Our technology eliminates the need for data migration, regulation, storage, excessive compute power usage and facilitates customer acquisition.
Our systems are compliant with the following regulators : GDPR, PIPEDA, CCPA and HIPAA.

What we do

We solve 2 main challenges in the greenhouse AI market

Accelerating end-to-end AI solutions

Most model development relies on a slow, manual feedback loop, which consists of physically speaking to the grower and manually entering data. TensoAI's infrastructure runs on-demand and live AI models at the greenhouse. Results are shared among local greenhouse systems, giving modellers a Data-Driven Feedback loop ensuring models are run at the highest accuracy.

Ag-Data Marketplace

Accessing private datasets in greenhouses is a long and challenging process. It involves legal services, contracts, servers, compute power, and time to migrate private data for AI training and development. Using privacy-preserving AI modelling, models are sent to the data at a large scale and train on multiple private datasets ensuring universal development. Access to this data is available through TensoAI's marketplace.

Our Focus

Run Models Onsite

Push the Models, not the data

Data Migration is slow and costly, increasing time and resources in development and consequently cost of acquisition. Our solution : run models at the edge where the data resides, in a secure and scalable infrastructure.

Data and Model Revenue Stream

IAAS infrastructure as a service

Pay only for the resources that devices (Codename: Elroy) use. We monitor the device resources (CPU, GPU, runtime) and charge the only appropriate parties. Our focus is to support a full end-to-end solution for both parties and increasing value together.

Validate AI models

Ensure models are ready for production

Most model developments do not have access to private datasets that would be used in production. With Tensoai's Private network, models are distributed across a wide variety of greenhouses within the desired crop type and dataset.

High-Quality training data

Speed up development by focusing on the models

8 of 10 AI projects fail due to low-quality data used to train and develop their models. TensoAI’s focus is on creating a high-quality data set for data scientists to create models using privacy-preserving tools.

How we do it

For Data Scientist

Select Plant Type and Data Type
Develop and Train with Privacy-Preserving tools and High-Quality data
Validate models using Data-Driven Feedback loops
Distribute your models across multiple greenhouses
Enjoy the Revenue Stream

For Greenhouses

Install Tensoai device at the greenhouse (Codename: Elroy)
Run AI models on greenhouse data automatically, anonymously and securely
Enjoy the Revenue Stream

FAQ

Tensoai's focus is on creating an infrastructure that knows how to protect the data. We do this by using our Tensoai Network to do remote executions at the data source. Remote feature engineering is achieved through Differential Privacy, noise budgeting and other techniques that protect the data. This allows a statistical analysis of the data without seeing the RAW values. If more transparency is needed, noise budgeting can be reduced to enhance the speed of development.

All data is protected through privacy-preserving tools.

All models are secured under a 2048-bit RSA SSL connection.
BIAS data comes at many different stages, from data capture to processing, and development. At Tensoai, we work with greenhouse systems to ensure the data being captured is at the highest-quality, at the very beginning stage. This ensures that the bias is not amplified through each stage during the development
Currently, we are focussing on 3 main crops, tomatoes, cannabis, and leafy-greens; and two types of data: environment (sensor) data and plant vision. These high volume crops provide quick feedback loop cycles.
We are always investigating new data points, and will be expanding our library over the next few years.
If you have or are searching for other data, just let us know.
No, Tensoai doesn’t sell data. We provide a gateway for models to be run at the greenhouse. Data is never copied. At the end of every script, the device's memory is cleared to ensure there is no remaining data trail. Additionally, the privacy-preserving tools used to develop models ensure total anonymity
Installation is super easy! Plugin both the power and ethernet cord (connected to your network) in an office or server room. We will handle the rest!

Elroy stays dormant until it is being called upon to run models. Running models is initiated by either a greenhouse system or a data scientist. An email will be sent to you before any model is run on your device.
Invoices and payment will be issued each month on the day of registration. Registration is only eligible once the device has been installed and verified (remotely) by one of our customer excellence technicians.

Payments are issued via PayPal, electronic banking or mail-out cheque.

Get in Touch!


Subscribe to our Newsletter

Be first to get all the latest Tensoai information!

Trusted System Collaborators