3 ms·

I did the same thing! I was a Quant at asset managers / hedge funds and then I started a data analytics company. Eventually I got into alternative data and then

I did the same thing! I was a Quant at asset managers / hedge funds and then I started a data analytics company. Eventually I got into alternative data and then my company was acquired by another data analytics company. I worked at the acquiring company for 2 years and it was actually very informative. I recently left to start a new company working on a mobile app. When I was working at the acquiring company, they didn’t really have a good place for me. I had more domain knowledge than anyone else at the company but I had less experience with even the basic etiquette of how software reviews work (I.e. GitHub pull requests). Coding as a quant and software engineering are really quite different. As a quant, you have to optimize for getting stuff done quickly at all costs. In software engineering, there’s more focus on testing, maintainability, and readability. Eventually I created my own role as a technical product manager and that worked out quite well.

platz 4y ago
what is 'alternative data' Is it like listening to nirvana when eveyone else is listening to michael jackson?
- dmurray 4y ago
  In the finance space, yes. Regular financial data is stuff like trading volumes and earnings reports. Alternative data is looking at satellite pictures of car parks to see which retailers were busy this quarter.
  - tomcam 4y ago
    Ok that sounds fascinating.
claytonjy 4y ago
> As a quant, you have to optimize for getting stuff done quickly at all costs. In software engineering, there’s more focus on testing, maintainability, and readability. I would have expected the opposite! I've never been a quant and have mostly done software/data science at smaller companies where there's very little focus on testing/maintainability/readability. I'm sure at bigger tech enterprises those things are stressed more. I figured a hedge fund would also care a lot about those things, at least testing, because a bug can cost a lot of money in a very direct way. How do quant shops get away with looser software practices?
- isotropy 4y ago
  Often you separate the feature exploration environment from the infrastructure that executes a strategy. You may need to rebuild the feature calculation before deploying it into prod, but if you try to provide the quants with solid enough tools so they explore the statistics of features and models without having to think much about the built in error handling, that can work well to avoid the rewrite step. You probably end up wrapping numpy and pandas with a very opinionated internal view of how the business handles the kind of edge cases your team sees in your data.

platz 4y ago

what is 'alternative data' Is it like listening to nirvana when eveyone else is listening to michael jackson?

dmurray 4y ago
In the finance space, yes. Regular financial data is stuff like trading volumes and earnings reports. Alternative data is looking at satellite pictures of car parks to see which retailers were busy this quarter.
- tomcam 4y ago
  Ok that sounds fascinating.

claytonjy 4y ago

> As a quant, you have to optimize for getting stuff done quickly at all costs. In software engineering, there’s more focus on testing, maintainability, and readability. I would have expected the opposite! I've never been a quant and have mostly done software/data science at smaller companies where there's very little focus on testing/maintainability/readability. I'm sure at bigger tech enterprises those things are stressed more. I figured a hedge fund would also care a lot about those things, at least testing, because a bug can cost a lot of money in a very direct way. How do quant shops get away with looser software practices?

isotropy 4y ago
Often you separate the feature exploration environment from the infrastructure that executes a strategy. You may need to rebuild the feature calculation before deploying it into prod, but if you try to provide the quants with solid enough tools so they explore the statistics of features and models without having to think much about the built in error handling, that can work well to avoid the rewrite step. You probably end up wrapping numpy and pandas with a very opinionated internal view of how the business handles the kind of edge cases your team sees in your data.