Yesterday, AI Trends covered the seed round funding of Almax Analytics. Earlier today, we caught up with their CEO, Balazs Klemm.
AI Trends: Congratulations on your seed funding. Can you tell us how much you have raised?
Balazs: We are very pleased with the caliber and diversified group of smart money investors that we were able to attract which is a large asset in itself for Almax Analytics. While we do not disclose the amount raised we can confirm that we were oversubscribed and that we will be entering our Series A round soon.
AI Trends: Given there are many new companies now focused on providing machine learning solutions in the financial markets, can you tell us why your plan is different and better than the others?
Balazs: I believe a lot of energy and money flows into areas that are measurable, rather than what is meaningful.
One of the advantages we have is that we are solving a problem that as a founder I encountered directly while running an investment fund – the inability for humans to read and process the sheer quantity of news available in order not to miss important events and relationships – and so we know exactly how to approach the problem and how the solution has to look like.
Talking with market participants we know there is a large schism between those that have the technology and those that have the domain knowledge.
We are closing this gap by having brought together a group of world class machine learning and AI as well as industry experts.
AI Trends: What software platform, analytics and machine learning software are you building your platform on? Can you describe some its technical platform features.
Balazs: The back-end is implemented using Java technologies. For sophisticated customized natural language processing LingRep technology is applied. As we scale it is planned to exploit Neo4j for semantic graphs, and Apache Spark for distributed data analytics. In the back-end web service interfaces cover all input and output functionality. The internal analytics pipeline is set up as highly parallel tasks that handle all processing and deduction steps. All external components used support distribution and scalability, and are locally integrated and addressed via web services. Each operational component such as web services and databases are defined as operational clusters.
AI Trends: Can you briefly describe the Almax process involved in capturing market data, both in structured and unstructured format?
Balazs: Almax Analytics is collecting and processing structured (i.e., fundamentals, quotes) as well as unstructured (i.e., news) information of several sources autonomously. Any input is connected via a stand-alone client that is operated in the distributed environment and transfers the information to a central data hub. Then, the real-time processing core takes over.
AI Trends: You state there are two main value adds to Almax. First is instantaneously taking the unstructured and structured data, combining it into a semantic model, which that itself will provide some ROI to its users. Secondly, taking that semantic model and processing it thru the Almax machine learning algorithms, which will then provide assistance to the trader, but is not expected to replace the trader as in an automatic trader. How much of this architecture will be available in the first release, and is the ROI that you would be able to state to prospective clients?
Balazs: While initially we start with a subset of stocks from the overall market we are covering in our system, the value for the prospective clients for each of those stocks remains the same.
For Analysts it will help them to intelligently gather and store information they would have otherwise missed and it greatly enhances their efficiency, for Portfolio Managers it is a great tool to enhance their existing Intelligence as they never miss significant news and they have all the data and calculations instantly at their fingertips to make an informed decision and for Traders and trading desks it enhances directly their alpha potential and trading strategies.
AI Trends: When do you plan on announcing that Almax is ready to go to market?
Balazs: We are building out the platform over the next 6 months with the aim to release then the first full version in a Beta-test phase.
AI Trends: Which specific vertical industries will you be focusing on when you launch the product?
Balazs: We will launch the product with US stocks and in particular with a focus on the alternative energy sector. We will then focus on scaling the system with an increase of the covered universe & topics.
AI Trends: How can Almax scale its software to provide solutions to many vertical markets?
Balazs: The general architecture and technology is applicable to any set where the extraction of unstructured data into a structured format and the combination with expert knowledge can generate insights. Examples include but are not exclusive: Healthcare, Compliance, Audit, Customer Relationship management.
Thank you Balazs.
interviewed by Eliot Weinman, Executive Editor, AI Trends