I write this as precautionary tale intended to Engineering undergraduate students interested in the areas of Data Science and Artificial Intelligence (initially given as a guest lecture at Mexico’s ITAM video-link in Spanish)
Data Science is supposed to be the intersection of Computer Science (hacking, understanding databases), Math and Statistics (advanced modelling) and Domain Knowledge (Finance, for this tale). Each discipline traditionally required its own career syllabus, and in real life is almost impossible to find someone who is proficient in all three fields.
In my experience, I have noticed that some people might be very strong in two out of three:
- Economists with Econometric specialties can solve Stochastic Equations on their sleep … and use VBA and Excel for Monte Carlo simulations.
- Computer Scientists can build databases that can handle tick by tick data … and be unable to compute the duration of a Bond.
- And Mathematicians assume cows are spherical and live in a vacuum.
Instead, the above diagram should be used as reminder of the skills we need to solve a problem and team up to fill the gaps.
Within Data Science Artificial Intelligence has gained popularity. For a more detailed review read my blog: AI in Finance: Cutting Through the Hype, but for the purposes of this story just keep in mind that currently AI is just a collection of tools that specialise in a very narrow task. I’ll borrow this graph for a moment:
AI has conquered the game space (Alpha Go), is great at writing like Shakespeare (take a look at the “Unreasonable Effectiveness of Recurrent Neural Networks” — go to the end of this blog for a sample code that uses Reddit posts), and can be used to create deepfake lip syncing videos (open source code or use wombo.ai — see art singing)