Machine learning, data science and AI solutions

Learn More
Image

Developing

Our Past AI/ML Projects

Analysis of Satellite Imagery

AI Platform for detection of objects in satellite and aerial imagery.

Car Damage Detection (CV)

System for car damage detection and assesment using image segmentation (Insurance industry).

Content Generation with AI

Training of GPT-2 on custom domain texts and deployment for internet marketing domain.

Sentiment Analysis (NLP)

Sentiment Classification of Structured/Unstructured Data, Aspect Based Sentiment Analysis (Clients are hedge funds and other financial institutions).

Dependency Parsing

Development of specialized dependency parsers for Aspect Based Sentiment Analysis project.

Food Classification (CV)

Food image classification using Convolutional Neural Networks.

ML classification of users on social media

Development of machine learning classification of user accounts on social media, for influencer analytics platform.

Recommender system

Developed a big data recommendation and collaborative filtering system, for client's platform.

Explainable AI

Built a framework to provide in-depth explainable AI and machine learning results for all of our projects.

Understand business goals

In the first phase of project, our goal is to understand the business goals of the project and to gain insights by applying data science on available data.

Define Roadmap

Based on results from first phase, we work with your team to prepare and define objectives of the project and detailed roadmap.

Build prototype

We gather the relevant data for training of the machine learning models. Implementation of the solution with API endpoints or as part of standalone solution.

Monitor performance

Monitoring performance of the machine learning model in operations. Further development and improvements if required.



Explainable Artifical Intelligence and Machine Learning

Ogranizations implementing AI and Machine Learning solutions are increasingly confronted with demands for explainability of their models results. This is driven both by regulations, e.g. Right to Explanation in GDPR as well as interest of other stakeholders - decision makers, employees involved using ML models, customers and others. We use advanced machine learning interpretability frameworks, including LIME-based and SHAP, to provide AI explainability for each of our models that we develop for our clients.



Explainable Artifical Intelligence and Machine Learning
image

Robo Advisor Solution

Alpha Wealth Robo Advisor is a platform for automated advisor and investment management, offering portfolio optimisation based on maximum drawdown, active management products such as dynamic asset allocation, financial planning and use of specialised digital assistants.

For an overview of Robo Advisor features please read presentation at Robo Advisor Presentation or see a video of app in use: Robo Advisor video

background

BittsAnalytics

BittsAnalytics is an advanced sentiment analytics platform for analysis of stocks, futures and digital assets.

It is used by many individual clients and global hedge funds.


Image
Image

Product categorization

Our product categorization solution allows you to automatically categorise all kinds of texts into categories (Google product taxonomy, IAB). Online stores can use it to improve customer experience on their websites - through better searching and filtering.

Our product categorization solution uses advanced machine learning algorithms for classifications - try out classifying texts by using machine learning restful API.

If you are interested in website and full path URL classifications, we offer Website Categorization and URL Classification API. We also wrote an overview article on Website Categorization API and also another one about categorization of websites - medium.




Image
Image

UnicornSEO

UnicornSEO is a collection of automated keyword research and niche research tools for improving your rankings through better insights on keywords, competitors and your niche.

Automated keyword research helps you find the keywords that provide most opportunity and that you should focus on, e.g. finding best skin care keywords.

You can compare and rank niches, see top domains in each niche in terms of search engine traffic, find their top keywords and compare then to your own domains. Excellent for content marketing data visualization.

UnicornSEO also includes various AI and machine learning tools for sentiment analysis of text, semantic similarity analysis based on sentence embeddings and AI Content Generation with GPT-2. It is used with another platform of ours for lead generation ai.

Visit UnicornSEO


Image
Image

Quantamental

Alpha Quantum Quantamental is an innovative solution for financial analysis, valuation and ranking of companies and a backtesting platform for quantitative strategies.

Solution allows clients to conduct financial analysis of companies, including generating sophisticated automated reports and provides versatile discounted cashflow valuation tools.

Quantitative strategies can be based on multifactor models, pricing models, rankings, DCF valuations, news analytics models or their combinations.

See Quantamental Presentation






Image
Image

Risk Management

Alpha Quantum Risk Management is an innovative solution for risk management in financial institutions.

Solution allows clients to conduct risk analysis through different Value at Risk methods, construction of stress test scenarios, support for derivative positions, pre-trade risk management and risk attribution.

The features also include limits monitoring, performance measurement, regulatory compliance and powerful reporting capabilities for credit risk, market risk and portfolio analysis.

See Risk Management Presentation


Image
Image

Portfolio Optimiser

Alpha Quantum Portfolio Optimiser is an innovative solution for portfolio optimisation in financial institutions.

Solution includes mean variance and mean cvar portfolio optimisation, powerful multiperiod portfolio optimization framework for backtesting and research of strategies, use of deep learning neural nets in backtesting of strategies, cleaned correlation matrix.

It includes portfolio optimization on cointegrated prices and alpha returns, powerful reporting and document generation capabilities. We also provide role as consultant in case you want to integrate portfolio optimization approach in your organization processes.

See Portfolio Optimiser Presentation


Technical stack

Machine learning models: Decision Trees, Random Forests, Support Vector Machines (SVM), XGBoost, Lightgbm, Logistic regression, Gradient boosting, (Ridge, Lasso) regression, Convolutional Neural Networks (CNN), LSTMs, K Means Clustering, Reinforcement learning, KNN, SVD++, PCA, Tsne, Content based recommender, Collaborative filtering, Real-time bidding. Word Embeddings (WordVec, GloVec), and others.

Frameworks: Scikit-learn, Tensorflow, Pytorch, Keras, Caffe, Surprise, Turi Create, Pandas.

Languages: Python, Javascript, C#, Java, SQL, CSS3, PHP, HTML.


Technologies we use