• C++ Programming for Financial Engineering
    Highly recommended by thousands of MFE students. Covers essential C++ topics with applications to financial engineering. Learn more Join!
    Python for Finance with Intro to Data Science
    Gain practical understanding of Python to read, understand, and write professional Python code for your first day on the job. Learn more Join!
    An Intuition-Based Options Primer for FE
    Ideal for entry level positions interviews and graduate studies, specializing in options trading arbitrage and options valuation models. Learn more Join!

Interview with Deep Q Digital: Building AI ready trading systems

Joined
5/2/06
Messages
11,750
Points
273
I'm very excited to see our coding extraordinaire @APalley, the TA for our C++ and Python online courses is interviewed with Deep Q Digital, an algorithmic trading firm in CT.
I've known and worked with Avi more than a decade and can say with absolute confidence that he is one of the best developers with a top notch communication skill.
He plays a crucial role in training many of our members to become competent developers with his unique style.
As the finance industry goes through huge shifts to AI, big data, etc, it's important to expand our discussion to cover crypto, AI and other FinTech opportunities.

Very interesting interview and I learn a lot

About Deep Q Digital: We're an algorithmic trading firm providing institutional-grade liquidity to the digital asset market. Our mission is to facilitate the advancement of fair, ethical, and efficient financial markets for the benefit of all.
 
Last edited:
Wow. A lot of helpful information in this video... I got caught up in it. The mentioning of kdb+/Q, of focusing on the technologies that have proved worthy in the past to deliver short-term wins (rather than working on a project for an entire year without immediate results), of minimizing risks at each point in the connection of different technologies (e.g. Python to kdb+/Q, and capitalizing on each tool's strength) was all helpful.

Amazing stuff to be honest...! Thanks for sharing.

Edit: this episode is quite useful too. Notably at the 36:40 mark onwards.
 
Last edited:
@APalley , I was reviewing this episode this morning.

You mention (timestamp 29:40): "kbd is really important if what you're going to be doing is going to revolve around time series data. I think the vast majority of quant work these days does revolve around time series data to a large extent. So then kdb is worth knowing. I've never seen a kdb class in any school.".

Silence ensues, and eventually Matthew Reid answers pensively: "Neither have I actually.", looking down.

Would the lack of available resources/expensive licenses explain the inexistence of such courses? Any thoughts? A new opportunity for QuantNet?

This would be quite relevant to what I would want to do in quant/macro modeling...
 
Last edited:
Last edited:
@APalley , I was reviewing this episode this morning.

You mention (timestamp 29:40): "kbd is really important if what you're going to be doing is going to revolve around time series data. I think the vast majority of quant work these days does revolve around time series data to a large extent. So then kdb is worth knowing. I've never seen a kdb class in any school.".

Silence ensues, and eventually Matthew Reid answers pensively: "Neither have I actually.", looking down.

Would the lack of available resources/expensive licenses explain the inexistence of such courses? Any thoughts? A new opportunity for QuantNet?

This would be quite relevant to what I would want to do in quant/macro modeling...
It's something we've considered, possibly at some point.
Seems First Derivatives from Newry, County Down, Northern Ireland (just down the road from my little house on the prairie) is active in this area.


FD is the parent company of Kx, which owns kdb
 
@APalley @marcusaurelius @Daniel Duffy

I found a kdb class!! Carnegie Mellon @CMU MSCF has a "Market Microstructure and Algorithmic Trading" elective option in the fifth and final 'mini' semester.

"In hands-on course assignments, you will utilize the industry-standard Kdb software to work with actual intraday transactions and order flow data."

 
Update, I found another kdb course. This is listed under Colombia's MAFN spring elective course options:
Screenshot 2023-04-27 at 2.58.45 PM.png
 
Very insightful podcast. Inspired me to check out kdb after I complete the C++ course...

I found some q/kdb courses from KX themselves, was wondering if anyone with experience could comment on how good they are:

KX Academy
 
Very insightful podcast. Inspired me to check out kdb after I complete the C++ course...

I found some q/kdb courses from KX themselves, was wondering if anyone with experience could comment on how good they are:

KX Academy
q/kdb is awesome. Very useful.
 
Not many resources out there though... I want to pick it up over the next few months. Any recommendations? I know there is Avi's textbook recommendation he had recommended.

Inhale Avi's textbook, plus their website. After that your best bet is to go out into the wild on github to see it in action. Set up a server similar to a SQL db type thing and start practicing! Integrate it with some of your python or C++ code and you'll be miles ahead.
 
Set up a server similar to a SQL db type thing and start practicing! Integrate it with some of your python or C++ code and you'll be miles ahead.
If you end up setting it a github would be really cool to share what you can do integrating with C++ or Python.
 
Inhale Avi's textbook, plus their website. After that your best bet is to go out into the wild on github to see it in action. Set up a server similar to a SQL db type thing and start practicing! Integrate it with some of your python or C++ code and you'll be miles ahead.
Makes absolute sense. Will do and show you the progress in a few months when things tone down at CMU!! 🤝
 
At work, I extensively work with time series data and have tried my hands on several timeseries db including influx, clickhouse, druid and timescale
For my company's use case, I decided to go with timescale. The main use case in which timescale's performance trumps other is aggregation of ohlc data for different time intervals. Plus, It is built on top of pgsql, so no need to learn a new query language, unless one is interested in learning q or influxql. I am sure it will pay dividend if one becomes a q-pro.

For the above mentioned use case, found this if exploring timescale - kdb+ with timescale
 
Last edited:
At work, I extensively work with time series data and have tried my hands on several timeseries db including influx, clickhouse, druid and timescale
For my company's use case, I decided to go with timescale. The main use case in which timescale's performance trumps other is aggregation of ohlc data for different time intervals. Plus, It is built on top of pgsql, so no need to learn a new query language, unless one is interested in learning q or influxql. I am sure it will pay dividend if one becomes a q-pro.

For the above mentioned use case, found this if exploring timescale - kdb+ with timescale
Very helpful thank you!
 
Back
Top