About Arcane Future
Arcane Future is an outsource company with the focus on analytics, big data and machine learning services for companies in the gaming industry
We help to develop voice assistants, build chat-bots, conduct AB-tests, increase players LTV, forecast churn rate and create big data storages.
In order you make decisions based on actual data we build charts by using data visualization platforms like Tableau, Power BI, Google Data Studio. We create KPIs tracking systems which automatically detect anomalies and send alerts via email, Slack, Telegram.
We setup process of conducting AB-tests, so you can quickly check hypothesis and your employees understand the concept of statistical significance and the difference between Bayesian and frequentist approaches to AB-testing.
We design and implement event tracking systems, develop events structure and help to build an appropriate cost effective and quick data storage on Google Cloud and AWS. So you can concentrate on product development while we are keeping data collection and management system up to date.
In order to analyze data from different sources, for example from MySQL, PostgreSQL, VERTICA, MongoDB, Google Analytics we write Python-scripts connecting all these sources. Our data gathering systems are scalable because we use Docker and Kubernetes.
We know how to use convolutional neural nets for detecting gestures and movements of human body, hands, fingers by processing video stream from the proprietary web-cameras. And how to train voice assistants and chat-bots to communicate with players.
By using neural nets and ensembles of decision trees for LTV forecast and churn users detection we can precisely gauge your CPI for every marketing channel.
We in help when:
Increasing the number of employees, which need data for decision making, and your developers don't have time and resources to make new tracking and unload data from the database
You need to conduct many AB-tests in order to maintain steady product growth and make tests show the correct results by understanding the math and statistics behind them.
You have difficulties with developing working KPIs system for different department and can't get the whole picture on everyday basis
You lose data because tracking system development can't catch with the development of the project.
Data in different sources are doubling and don't match.
You want to use machine learning but it isn't clear where is the biggest potential for applying these technologies.
We bring the most benefits if turn to us before the project start. So we develop data gathering system, create charts with necessary metrics for employees and prepare project for using a machine learning.
We know how quickly check the predictive power of any dataset, by using prepared pipeline with the standard models. This allows quickly estimate the basic level of accuracy from what project is starting
In case you have a lot of features in the dataset and this increase model training time we apply dimensionality reduction techniques like PCA or encoders, by using these techniques we can preserve most of the dataset information and significantly speed up training and inference time.
In order to keep your product code safe and independent we integrate our solutions via micro-services and by using hi-load architectures our micro-services keep the response speed under 100 ms.
Make quality work is important for us so we devote a lot of time polishing details. But we deliver work on time because our working schedule based on short and effective iterations (a.k.a. sprints).