Throughout 2021, we’ve had a lot of ups during our growth journey:
Starting a consulting contract with BCM Partners to automate their discretionary analysis of the forex market
Transitioning this aforementioned agreement into a sub-advisory model based on our own proprietary FX trading strategy
Bringing on additional talent: James as a senior Quant Trader and Jonathan as a business strategy advisor
At the same time, we’ve also had equally as many downs:
Rushing through the creation of an ML model that was designed to predict short squeezes – this project ended up failing before we could even act on its insights because the model incorrectly learned the importance of certain features as a consequence of the limited amount of training data and examples available at the time of development
Taking on more responsibilities than we could reasonably handle – not only did this push some of our timelines back, but also reduced our overall quality of work during some points in time
Launching an incomplete version of our tools platform after months of development work over the summer – we have since pivoted away from this opportunity for the foreseeable future
However, we’ve been able to learn from all these experiences, we’ve recalibrated and by focusing solely on our algorithmic trading efforts, we’ve been able to brainstorm the core principles behind a new framework that will revolutionize the way quantitative trading is done. These ideas are all wrapped into a project called Chirp, which is designed to solve three major pain points:
Bridge the gap between strategy research and deployment through a universal interface that works across different execution environments from backtesting to live trading
Allow for multiple financial instruments of different classes (including options and other derivatives) to be traded simultaneously with information being shared between different asset controllers
Incorporate modular components that allow for numerous lenses of market analysis to be completed before trade entry
Chirp will also have additional features baked in such as a parameter optimization tool that uses reinforcement learning and other advanced techniques to increase portfolio returns, a sophisticated risk management engine that identifies potential tail-end risk, and the ability to improve trade performance over time by learning from past trades. In order to acquire the data sources necessary to develop a prototype version of this framework, we have recently closed a $30,000 pre-seed round with a close friend: Mohammed Nasir. Stay tuned for more information over the coming months as we turn these ideas into a production-ready system able to manage a large AUM.