Vega Intelligent Solutions: Advanced Human Interactive AI


The project described here was created over the course of a painstaking Journey of research and discovery and involved the ideas of many different people from all around the world. It is important to understand that much like the premise of this concept the very reason any of this is possible is because the original Axiom is true that the whole is greater than the sum of its parts. Many of the greatest technologies that we take for granted today are amalgamations of pre-existing technology that has been in practical application for the majority of our adult lives. To that end, I can’t stress enough how the potential of this technology has the same ability to impact every sector of technology and every aspect of our daily lives.

Birth of a star

Initially the concept for Lyra began as my own research into algorithmic trading technology. Keep in mind I’m a junior developer when it comes to full stack development, though I do have over 10 years of Python experience mostly involved with scripting and modification of pre-existing code. I begin day trading via various crypto platforms, but I quickly found the process tedious as cryptocurrency does not move at the volume I desired. Not wanting the red tape of brokerage platforms and given the fact that I deal primarily in cryptocurrency, I eventually found a few platforms like One Broker and SimpleFX that allowed me to use Metatrader and trade Forex.

It wasn’t long until I discovered that the optimal strategy for trading Forex is algorithmic trading, and there’s a lot of pre-existing solutions for this some of which I had a lot of success with. At the end of the day though my success came down to optimization, and optimizing the parameters for any form of algorithmic trading takes time and CPU cycles, to some extent machine learning is already built into a lot of the platforms that are offered and utilize these strategies. However after testing a lot of the different platforms, I found myself sticking with Metatrader and its built-in optimization system as it worked the best for cryptocurrency trading.

Given the constraints of the tools I had I immediately sought out other resources for algorithmic trading to understand the fundamentals of the Price Action System trading and how to determine the best price actions at any given time-based on historical events. Naturally this led me to two forms of machine learning called Q-learning and MF learning. Using the resources I had I was able to create a trading bot that performed relatively successfully, but I wanted more — I wanted to create a system that stripped the mystery out of algorithmic trading and removed the learning curve that took me literally more than a year to build upon the foundation of data science and algorithmic trading required to even make a little headway.

A constellation of doubt

I’m the first to admit that I’m not a data scientist. The research and software that I progressed into became a learning curve far above my education and training, it was immediately evident to me that I had to call in backup. This is when I approach my good friend Brian Carter. He and I had work together in the gaming community for many years, and I found him to be an extremely competent and capable individual. I also happen to know that he was a great software developer and an amazing software architect. Both professionally and personally we’ve deployed solutions and brainstormed different ideas that would blow people’s minds. I proposed the idea to him over lunch one day at a little bar inside of a Dallas hotel called the NYLO, and he was immediately captivated by the idea. He took some time to chew on it and process the possibilities, and it wasn’t long before he made it his own, in fact very name of the project, Lyra and the AI called Vega, was his idea. The name itself pointed to a constellation where Vega (also known as Alphae Lyrae) is the brightest star within it, which translates to the core of our machine learning brain in the Lyra Neural Network.

To begin building his team, Brian sought out a data scientist, professor, and software developer he’s worked with in the past, Ryan Schreck, and they got to work researching and planning a framework for this technology that began evolving past the concept of even a simple algorithmic trading platform. Before my eyes, and in no time at all my concept for something as simple as predicting trades evolved into a human interactive intelligence that was modular in design and capable of performing tasks in almost any industry.

Inside the brain

Our intentions are to create something that is to utilize a Deep Reinforcement Q-Learning and many other techniques to create original algorithms and optimize existing ones that are currently being used in the market to make the very best possible decisions. Enable for us to accomplish this, we need to feed the Vega AI as much data training information as possible at a very high frequency. This involves creating a neural network that is capable of processing this data with the latest technologies. We can do this by establishing a multitude of data collection clusters through API service calls that backfills historical data and obtains real-time data for cryptocurrencies and other world markets with credible full articles, twitter feeds, forum threads, and market trends that the AI can process, analyze, and train itself to create decision trees and correlations between all resources. In doing so, the Lyra network will be a fully automated machine learning network that users have available at their fingertips.

Human interaction

Natural language processing (NLP) is the cornerstone in quality for artificial intelligent solutions. It allows the user to submit free text searches and analyze documents for keywords within the AI’s brain to find the most useful information. It is important for us to create a superior user-experience and interactions between the investors and Vega directly. There is a very well-known NLP algorithm called Gradient Descent which is highly effective in neural network optimizations. Using this algorithm we want to give the user the ability to control their investments and find the information by speaking freely to our AI platform. Using natural dialog on integrated platforms such as Slack, Skype, Discord, etc. and our own built-in application with a sleek and intuitive user interface. If I want to ask Vega what is the best investment for me and automatically trade I simply do so and in return Vega will give me several parameters or questions as a response to help make the best possible decision for my portfolio. Many factors will play a part of that such as the investment amount I’m willing to spend, how aggressive I want to buy and sell, how many different investments I want to make, etc., so the AI can use that information to build that portfolio and transaction plan.

More than a single star

By no means is this the end of our journey, the applications and potential of this technology are truly limitless and if we achieve the goals that we’ve set forth in our roadmap and are able to truly manifest the destiny that we are passionate about, there is no doubt that the future will have a very different shape than the one imagined today. Consider the perspective of the future as seen from the 1960s in the classic movie “2000: A Space Odyssey”. The future we imagined today is extremely dissimilar from the vision they had back then and one could argue that it’s more realistic, but Quantum leaps in technology do happen from time to time and this is one of them. imagine being able to ask your television who the actor on the screen is, imagine being able to tell your computer to automate tasks in the most simple terms and have it truly be able to take over the work and produce satisfying results with little-to-no configuration, imagine if you will a world where Logistics and planning solutions are solved effortlessly, every day by an artificial super intelligence powered by you.

Welcome to the future, built by Vega.

Article written by:
Notorious from
Brian Carter from