Automated trading systems

Our clients have invested in multiple trading strategies with Sharpe Ratio over 2.0 in both traditional and cryptocurrency markets. Strategies were developed with the use of state of the art machine learning algorithms, NLP and market insights. Using portfolio optimization enabled our algorithms to be profitable in both bullish and bearish high volatile markets.

Our team has identified more than 200 technical, commercial and market risk sources in the cryptocurrency market and developed risk management strategy which allowed us to minimize losses in market force-majeurs.

It includes such software as real-time risk management system, monitoring system, algorithms administration system, data mining system and "market sandbox".

Python / C++ / C# / Machine Learning / Time Series Analysis / Mathematical Modelling / MongoDB / AWS / Azure / Docker / CI/CD / REST / Websockets

SemanticForce: Aspect-based sentiment analysis
The value of the project is cutting costs and optimizing time and process for social media analysis with optimizing routine work for more than 30% of the staff with fully automated AI pipeline.

The solution was the state-of-the-art NLP system for fine-grained aspect-based sentiment analysis, topic modeling, and text categorization for different languages and data sources including automated machine learning pipeline.

Our team have leveraged the latest developments in deep learning-based NLP for text vectorization and generation to deliver one of the best solutions on the market.

Python / NLP / Deep Learning / PyTorch / Tensorflow / SpaCy / NLTK / StanfordNLP / gensim / REST
Ilogos Europe: player analytics model
We provided a complex solution to our client who asked us to solve following business goals: deeper player's engagement due to the customized gaming experience and game designer's time economy by giving him a "playing pulse" of each player in real-time.

The technical solution itself was a novel mathematical model on the intersection of data science and cognitive modeling.

The model provides a single number, that is a predictor of churn, in-game purchases, overall performance and satisfaction with the gaming process.

Python / Machine Learning / Mathematical Modelling / Deep Learning / Keras / Tensorflow
MLVCH: style transfer for photos and videos
Our client became one of the earliest adopters of the artistic style transfer technology, that is "repainting" given photo with an artistic style of a particular artist or a photographer. We have adopted and implemented technical solution, that became one of the first consumer apps on the market both for web and mobile platforms.

Further, Mlvch was extended as a B2B product that involved video style transfer for marketing campaigns and another B2B product for automated photo retouch based on a given artistic style.

Python / Matlab / Mathematical Modelling / Web Development / OpenCV / Tensorflow / PyTorch
Image logo retrieval

Logo retrieval is needed for customers of the end client to obtain all the visual mentions and appearances of their products in the web. The main purpose of the project is cutting costs and saving time and optimizing process in finding logotypes overall in the web to analyze the context of their appearance and the audience with the use of AI.

The solution was the computer vision algorithms for detecting logotypes of a particular brand on the web-large scale.

Python / Deep Learning / SSD / YOLO / RetinaNet / Tensorflow / Keras / REST
Ukraien, Kiev
Cupertino str. 132
+44 330 027 2146
International House,
64 Nile Str, London, N1 7SR,
United Kingdom
© 2019 Neurons Lab LTD. Registered in England and Wales
Company Number 12265479