Machine learning datasets. Curated by humans.

Knosis is a marketplace for curated human input given on machine learning datasets and data streams. We enable your internal team or your crowdsourced team to capture, tag and generate training and evaluation data sets for your machine learning and computer vision assignments.

Implementation of Services

Our wide range of ML solutions is now at your fingertips. We deliver end-to-end solutions for smart surveillance, digital proof of identity and data management for machine learning while providing full support throughout every stage of the project.

long-logo-envisae-mark Surveillance

Maximize efficiency

Boost the efficiency of your surveillance teams and create a custom retail experience for consumers with our end-to-end solution for smart surveillance and face recognition.


Trust Bridge

Avoid online identity fraud

As user identification is shifting online because of the large number of digital services, our solution helps you detect and prevent identity fraud.



Reliable data for reliable machine learning

Our data-centric product manages the data that is used for training machine learning. Its role is to bring quality through data accuracy, accountability, and completeness.

What sort of data can I use Knosis for?

Knosis is currently focused on computer vision datasets, which have to do with

  • Tagging object in photos
  • Tagging objects in videos, short clips or live photos
  • Identifying pictures which have been edited, generated or falsified digitally
  • Rating objects by quality or preference
  • Challenge time! Tag objects in photos using polygon tool

    Artificial Intelligence applied to your industry!

    Find out solutions that will offer a strong competitive advantage.

    What are our technological capabilities?

    Computer Vision and Visual Machine Learning

    In order to build Visage Cloud, our face recognition and smart surveillance solution, visual machine learning was at the core of our research. With real-time notifications and face detection and recognition, security staff can take quicker, informed decisions, boosting their overall efficiency with the help of our security and surveillance solution.

    Forecasting, Scoring, and Diagnostics

    With the continuous development of IoT Apps, comes a great level of information and data. Machine learning brings the possibility of a more complex contextual analysis to the table, adding in details such as environmental factors and climate changes, for instance, in the case of crop management. Through solutions like enten , our team managed to put the data
    to use by turning field measurements into actionable insights.


    Natural Language Processing

    A machine finds it more difficult to generate a text that conveys emotion as well as an actual writer. However, Natural Language Processing has advanced within machine learning and while it can not yet convey emotion, it can correct your spelling, accurately translate chunks of text, or even detect patterns to predict a certain type of human behavior or feeling.

    Business Process Machine Learning (BPML)

    When having multiple client applications (for credit, mortgage or even for a job) machine learning algorithms will use previous similar case history to sort them out and determine which are the ones with the highest possibility of success.
    This helps employees take the best decisions by prioritizing the most relevant cases and offering suggestions of the most likely outcomes.

    Why is machine learning the better solution?
    If up to now, computers only executed the instructions given through programming, machine learning brings meaning into play. Based on real-world data, the machine is able to attribute meaning to its end objectives by learning from previous outcomes, or by following a set of specifications.
    From analyzing and identifying patterns in large amounts of data, down to facial recognition,
    ML algorithms can easily help eliminate issues of bias, representativeness or privacy by
    keeping track of the data used for training.

    So no matter the industry that we adapt our solutions to, we make sure that we bring quality
    and efficiency to your teams and your business.

    IT Automation

    Automatic scaling, incident detection, proactive investigation and predictive diagnostics are all tasks that are meant for machines rather than humans. It is for this reason that we designed our infrastructure layers and service layers to have mutual visibility. Our built-in autonomous monitoring, scaling and security services keep the humans (i.e. developers) in the loop, but they perform all the retrieval and consolidation of diagnostic data behind the scenes, before human attention is involved.


    Human Potential Development

    In the rush to learn new skills we can no longer afford to await our partners and employees to learn by repetition and memorization of abstract best practice. Knosis allows them to learn by doing, by trying and erring in a safe experimental space. Leveraging the latest advances in natural language understanding and generation, collaborators will be able to acquire hands-on digital and critical skills, in a gamified manner that is considerate of their attention patterns and individual ability. 


    We believe that modern knowledge workers should learn by doing.



    As the mix of information capture by the IoT infrastructures of our agriculture partners grows in both breadth and depth, we have developed novel ways to integrate on-the-ground multi-spectral imagery with measurements from a distributed mesh of sensors in order to give agriculture experts a birds eyes view of crops and greenhouses, both in space, time.

    By leveraging the DeepVISS-Sense measurement standard, Knosis services can mashup data points from various sources, with various hardware and software architectures, while minimizing the amount of coding needed for integration.

    Actionable insight helps farmers reduce pest and disease, while reducing pesticide use and maximizing expected yield.


    Legal & Compliance

    Using the Knosis capabilities for Document Understanding and Semantic Character Recognition you can leverage a blend of human and machine for transcribing, sifting, sorting, cross-relating and cross-referencing legacy documents and data sets.

    Through a wide array of machine learning and adaptive intelligence algorithms and methods, including SWT (stroke width transform), GPT-3 and variational semantic (auto)encoders, Knosis gradually performs the handover of the menial tasks from experts to machines, so that experts have more time to dedicate their attention to those matter truly worthy of human insight, intuition and depth.


    We believe that modern knowledge workers should always get a change to do when they learn and learn when they do.

    Discover how AI works in the real world!

    See how Smart OCR, biometric login or IoT systems integrating AI work for your everyday performance


    The entire stack of our solutions build over OpenAPI and in full compliance with DeepVISS semantics.

    For interconnect and exchange opportunities, the Knosis Exchange (KNX) has full interoperability capabilities with existing OpenRTB and IAB-compliant endpoints and vendors.


    Knosis Data Management Platform (DMP)

    Managing augmentations for machine learning and data curation sets or streams (image, video, audio) so as to allow out-of-the-box auditing of performance and robustness for various existing models in your portfolio.

    For customers, this turns the domain discovery and definition phases as easy as dragging and dropping data sources and taxonomy elements.

    Knosis Demand-Side Platform (DSP)

    The Knosis DSP is useful for scheduling and prioritizing several augmentation pipelines between machine learning workloads, so as to maximize efficiency and throughput and lowering cost. This allows for a simple definition of domain models, taxonomies, as well as managing the storage requirements for growing data sets. For added efficiency, processes can be divided into several stages, so as to allow a dynamical split between human collaborators and programmatic components.

    Knosis Supply Side Platform Console (SSP)

    The Knosis SSP allows you to leverage your unengaged (“benched”) workforce by connecting as a supply-side collaborator to our network of digital nano-assignments. By optimizing a trade-off between knowledge utilization (i.e. using current skills) and knowledge acquisition (i.e. learning new digital and automation skills).

    Knosis Deployable Exchange (KNX)

    For customers who already operate high volumes on both supply-side and demand-side, we are also licensing a deployable version of Knosis Exchange (KNX). This allows for more efficiently internalizing operations at scale and it is a recommended option for enterprise customers with more than 5 million HCR per month. 

    recent Post
    Be part of our team!

    We will not give you a superhero speech or try to lure you with over the top hyped statements. All we can say is that we are a small, dedicated and experienced team who is looking to welcome more colleagues to teach the ways of AI to and to learn new industry developments with.

    Just as a heads up, we are looking for those who are familiar with  TensorFlow, Theano, Keras, PyTorch, Spring Boot, Python, Jupyter and Docker. So if you are passionate about AI and have experience in developing machine learning solutions, all you need to do is hit the “Apply now!” button below and fill in the form.