I'm a software engineer/business owner with 20+ years of experience. I've been doing a deep dive into artificial intelligence, neural networks and trading. I’ve been inspired by the videos where AI can outsmart and beat humans in computer games. I’ve read all the latest AI research papers and based on that, I wrote my own Neural Net training and execution software for the financial markets. I grow neural nets on a substrate and evolve those using genetic concepts from nature – DNA crossover, mutation, speciation, surviving of the fittest trader where each trader’s fitness is measured by how consistently it can make profits in the market. I'm doing my tests in the cypto markets because L2 data feeds and API trading access is readily available. I’m seeing some great results when I apply my genetically evolved nets on the order book and order flow. However order execution on crypto is bad, exchange servers can’t handle high load periods, and slippage and high fees are killing it. I’m looking for a partnership/collaboration with a trading firm that has access to any kind of market with low latency and low fees. As a retail trader I’m climbing a steep uphill. My system can take in order flow and can find microstructure patterns and can extract very profitable trading signals. The training of the neural net is run in parallel on 96 CPU machine for several days while I’m playing the role of “God” directing the evolution based on what performance I’m seeing. I’m well seasoned, and know all the pitfalls of curve fitting etc... My system can generalize and can extract alpha and have huge potential. I’m looking for low latency/low fees execution, which is not available in retail. I’m based in Manhattan, but I’m doing all this in a vacuum without talking to anyone. I’d appreciate any contacts with market making firms, FX or HFT firms. See some of my tests at https://www.youtube.com/channel/UCwvxSo9dsjxKrHPdcssvNWw Thank you! Rob the Quant
You’re looking for money. Nobody is going to give you any until it’s applied. You’ll need a couple years in the market.
I'm not looking for money, I'm looking for an access to low latency/low fee execution market, I have an income from my business, and can fund trading myself. In your opinion what's so unrealistic? The screenshot you posted is about 3 hours of trading only going LONG when Bitcoin is going downhill. Here's screen as it appears now - about 15 hours of live trading. see Live updates at dnatrading.info
Lease a badge and colocate. You'll have people throwing rates at you. MMing firms and HFT = generally sell-side. So either you're looking for a gig or you looking for a good rate and low-latency. That's easy. The sell-side won't look at your algo. So that leaves the buy-side. Do you have any experience outside of coding that would allude to you understanding anything about cap-constraints? If your BTC thing can earn a few K per day then I suggest you let it fund you as your only overhead is the microstructure.
the hang up i have with ai (over 30 years experience before the term quant become popular) is that ai is actually a continually self optimizing thing. example we have price anomalies or shocks in the data occasionally and ai can not sort that from the rest of the data. you have to pre sort the data before it's headed to the ai otherwise ai will create conditions to take that shock into account. which will affect future output, basically ai can never stop learning on dynamic data. it's great for dead static data but such as markets forget it, just another buzz word to put on a resume. there is absolutely nothing that ai can accomplish that naked price action can't out perform.
This forum is interesting... lots of strongly opinionated people with different points of view. I guess trading is based on disagreement, that's why. destriero: why did you say: "The sell-side won't look at your algo." ? I'm doing deep research to apply this technology for market making. People don't really understand that what I have is NOT one single strategy but a method of building various strategies that self-adapt to the dynamics of given market. I don't tell it buy or sell when x happens. I feed it order book data and it figures out by itself what to do to make steady profits. Of course it will try to curve-fit random data, but that's where I come in and guide it in the right direction so it generalizes and profits on future data. This is not a weekend project, I spent solid 6 months 10 hour a day to make this happen.
wow a whole 6 months of development - congrats lol as far as jets i would recommend a challenger make sure you get one with plenty of head space or it makes you look cheap when you visit monaco. nothing worst than hunkering down in the isle of a cheap jet.