We work to make it real!


Crossing Minds uses cutting edge machine learning techniques to extract customer data and provide recommendations tailored to any brand.

Our technologic approach is built on three components: probabilistic matching, machine learning embeddings and improved recommendation engines.

Machine Learning Embedding

Our Work

Transforming raw data into features or embeddings is often part of the process that most heavily involves human time because it is driven by intuition. Recent developments in deep learning and automated processing of images, text, and signals have enabled significant automation in feature engineering for those data types. Yet, feature engineering for relational and human behavioral data remains iterative, human-intuition driven, and challenging, hence, time-consuming. More importantly, our objective is to be able to do so in a domain or brand agnostic manner.

And in Simpler Words?

Crossing Minds is the first and only company in the world to provide a full encoding of consumer’s behavior and taste, without ever putting their privacy at risk.  

We are able to build a “digital DNA” through the embedding of information gathered through cross-platforms by creating an anonymous mathematical representation of a consumer that contains all their tastes and behaviors.

Our Modest Research on this Matter

An embedding is a low-dimensional vector of continuous numbers learned to represent relevant information in a compact and convenient form. Embeddings are designed to reduce the dimensionality of both categorical and continuous variables and meaningfully represent categories or features in the transformed space. 

Embedding is the future when it comes to reducing the dimensionality of categorical and continuous variables. The challenge has always been deciding which clusters, classifications, and actions are significant and turning this data into features. 

What do Other People Typically do?

Not much.

Deep Learning Recommendations

Our work

As a team of researchers with backgrounds at top institutions, we have developed a set of content recommendation engine that performs 20% better than the existing state-of-the-art and is the first application of group recommendation technology (recommending bundles of content to multiple people simultaneously) in a live consumer product.

And in simpler words?

The concept behind "deep learning recommendation" is similar to the reason why you are asking for a movie recommendation to a friend instead of a new acquaintance. Your friend will consider much more than your historic of movie to recommend you something such as your taste in music, in literature or your behavior. All those small part of information will provide with insights that are valuable in order to guess what movie you'd love to watch next.

Our Modest Research on this Matter

Recommending new items is a common challenge for businesses. More generally, any item without enough history of interactions from users won't be recommended by traditional methods. A novel solution to this cold-start problem is to extract deep information from text or image data using advanced natural language processing and creating a neural network of associations. Deep Content Extraction allows our software to recommend items no one has interacted with yet because the algorithm is able to understand the genre of the content by automatically extracting information from items such as a movie poster, synopsis, or reviews.

What does other people typically do?

Other solutions commonly patch their recommender system with hand-crafted rules to diminish the cold-start problem. However this requires expensive iterations and does not scale.

Probabilistic Matching & Identity Graph

Our Work

Businesses have a lot of available data scattered across third-party tools and contractors, such as Google Analytics, CRMs, and data from partner retailers. However these data sources do not communicate, and users are not identified uniquely across the data sources, which makes it impossible to leverage all of the available data.

Crossing Minds can leverage the overlap over the information collected by all these data sources to effectively reconnect the fragmented data. Using our proprietary matching algorithm and machine learning expertise, we will create an Identity Graph that will unlock the potential of all the scattered data already available.

And in Simpler Words?

Crossing Minds can help you connecting users across all your touch-points, removing duplicatas and generating end-to-end user journey across your platforms. This can be done even without any obvious overlapping on the information about your consumer across really distinct databases.

Our Modest Research on this Matter

Our user matching solution is completely unsupervised: ground truth for such matches is unavailable. Most methods in matching in academic literature, as well as industry solutions, require ground truth and labels.
Our algorithm utilizes carefully calibrated distance metrics and probabilistic models to provide matches across data sources, even without available ground truth.

What does other people typically do?

Most other solutions do not include any probabilistic matching to compute their identity graph, but mostly rely on deterministic matching. Deterministic matching algorithms are limited in the sense that overlapping between databases is not always doable, if not rarely, leading business to leave out databases such as CRM or Marketing Campaign out of the equation.

Interested in trying our Recommendation API for free?

We are very exited to share our newly released Recommendation API, which businesses can utilize to offer their customers the best recommendations possible.

If you are interested in being among the first to try it out, please answer the interest form.Our team will reach back to you as we are processing the flow of applications. Thanks!

Thank you! Your submission has been received!
Oops! Something went wrong while submitting the form