Linux Engineer wondering how skills apply in trading.

Discussion in 'Trading Software' started by berti, Feb 4, 2013.

  1. berti


    This may be a strange post for this forum, so forgive me, but as the thread title implies I am a bit out of my element being here in the first place.

    What may be more strange is the premise. I am a long time Linux developer and coder, currently working in a CIO-SysAdmin capacity in Manufacturing. My technical background spans the gambit with the core being in software development, across many languages and application-types, focusing on solutions and optimization. It came to my attention in a pub talking to a trader-type here in Chicago that a particular strong suit and interest of mine, that being data handling through text processing methods (bash, sed, awk, perl) to facilitate third-party integration and data routing/reporting, is a sought after skill set in your industry.

    Now I must admit to my ignorance in that while I understand financial markets and their movements, even (perhaps more so) in terms of statistical analysis, the types of data and patterns developers are dealing with and even the leading methods developing and deploying trading software are barely on my radar, for whatever reason.
    (Perhaps because my technical background is steeped heavy in the world of open source knowledge software, and the culture of individuals that are usually involved in such types of software communities are rarely so focused on applying their skills to the trading industry, pursuits of zionist money-mongers etc... more of an academic approach).
    That being said, as I looked in to this notion that text mangling geeks like myself might be paid oodles to awk the obvious out of some ongoing data-streams based on some trader-dude's obsessive analysis, I found myself asking more and more things I could find no answer to: "What example is there of this data that needs 'analyzed and optimized' for someone outside of the industry? Where is the comprehensive knowledge source that talks about applying {the aforementioned Linux skills} to optimizing micro-trading processes? Hell, what's a f#&$*ing intelligent resource at all that talks about this area of trading @ all?"

    Did I miss something in college? Were there undercover Linux geeks in the MBA finance program and I was too busy writing papers about how 'personal values and aesthetic interest affect scientific outcome' for my electives? Fill me in folks. If this is totally out of line, by all means disregard. But any links to helpful information or thoughts for a savvy but slightly from-a-different-industry geek that wants to know are much appreciated.
  2. shemz


    What your pub buddy was referring to is a popular cult in modern finance, called quantitative trading. Read the wiki article for a full blown glorified account from an academic perspective. Simply put though, they apply statistical methods to recognize patterns in market history that can be exploited with high precision (some also apply these method to create synthetic trading instruments, but the underlying philosophy is same). Side to the quants and traders also work coders and data analysts, whose primary task is writing scripts for string manipulations (R, Awk, Python, Perl are common here). Then there are software developers who make the trading systems work (C, C++, Java, C# mostly). Some trading firms also use exotic functional languages like OCaml and Haskell. You can find a lot of jobs advertised in these areas.

    The big question here is, as your topic title suggests, if there is any correlation between your coding skills and/or technical knowledge and the probability of becoming a successful trader. The answer is, zero. As a matter of fact, the more scientific you are, the less is your probability. The deterministic mindset carefully crafted by your professor and then your employers is what make most men of science completely blind to the randomness of the real world. (Fyi, I understand it rather well as I gave up my engineering career to trade full time, but I had to face psychological challenges in the beginning).

    Nonetheless, if you are not interested in actual trading, then there are a lot of positions in trading firms and banks which will pay much higher than tech firms. But since you are not the only one to know the benefits of working in finance, expect a very strong competition. Alternatively, you should stay with your present work, and start paper trading on your own to learn how the markets work.
  3. nitro


  4. As you say, using sed and awk in this manner is obvious. If that is the skill you are selling, you will not be paid "oodles" for it from "zionist money-mongers".
  5. heech


    What I read was "third data party integration" and reporting. That would make more sense from a pure text point of view. Various trading platforms/tools are still reporting positions + trades in non-standard formats, and having something to integrate that together is always nice.

    But no, I don't see that as a path to immense riches. If you want to make money in this industry (just as in any other industry), you need a clear/visible path to ways where you directly produce money.
  6. I've never met a "CIO-SysAdmin" before... that must be a rather small manufacturing company for you to set the direction of the overall technology in use and have to manage as well...

    anyhow... systems and application integration, sure... you have a place with your current skill set.... quantitative trading? not so much...

    some trading can be a science... while other trading can be an art... quants will focus on the science... desk traders will use the quant models and their "art" to make the trade...

    just my opinion...
  7. berti


    Thanks for all of the replies, and no it's not necessarily my goal to actually become employed in the industry, but as I said I've been developing software for some time and I become curious when someone sets me on to something in my field that I am not completely aware of.

    -- To the gentleman that's never met a CIO that also puts hands on the actual operations they set in place and oversee in manufacturing - really? Aren't you fancy. - fwiw, a job title is supposed to describe what you actually do, not what you want people to think your position consists of. :p - but to entertain your cuff-links - it' is a relatively small factory by industrial standards, - ~100 employees, however for our industry, we are the biggest factory in our metro area (Chicago). With only about 30 computer-end users in house, a few remote sales, and only a handful that aren't strictly data entry machines on the manufacturing floor, so there isn't too terribly much of a system to administer, so I do it because its too little work to hire in house to do specifically and it saves money.

    Particularly useful and interesting was the document on parsing URLs and regular expressions, this is closer to the core of my curiosity in terms of those types of skills in your area of works. Thanks.
  8. berti


    Shemz - Very useful explanation. Forgot to mention that. Thanks.
  9. I hope you didn't get offended, wasnt my intention at all... to me titles are worthless... I have met lots of CIO/CEO of 5-10 employee companies... or even better, MD's of the same size co's... and I chuckle because they think title means something... heck, i am CEO/CIO of my own conglomerate then... :)

    no title, doesn't describe what you do either... not always anyhow... a "CIO" setting direction/managing a small ecosystem on a 100 employee manufacturing company can't scale that skill set implied by the title to a 50K server, 6K employee and 3B IT budget co.. but anyone can argue with what I say... and it wasnt about what I want it say but what it stands for and means...

    in any event, I merely made an observation about how you described your position... no judgement was being made... as I said, I never met a CIO-Sysadmin ... but I have met plenty of system managers... (not meaning sysadmins)...

    now .... getting back to the quant trading stuff.... I dont know about unix native based regex tools, but the desks I do designs for, many of them use purposeful tools for the effort of identifying patterns within the time series... tools such as OneTick, Apama, etc... that can be used for prototyping and deploying models at the same time to ones SOR and OMS....

    PS. I do wear cuff-links... but not always... I prefer business casual and I only wear suits when I have to meet with the business MD's... or if I am interviewing... oh and btw, kudos for keeping manufacturing alive and competitive in the USA...I mean that... :)