How IntoTheBlock uses machine learning to analyze crypto assets

0:00I'm Ashley Addison from event chain for investmentpitch media and FinTech news network and today on blockchain interviews we have Alfredo torero the CEO and co-founder of into the block Alfredo thanks so much for being on the show it's a pleasure to have you here today thanks for having me you're very welcome if you could please kick off the interview by giving a bit of background

0:20into how you got into the cryptocurrency space analytics and how into the block was formed sure so I started in crypto around late 2016 i was actually i got into groups actually through etherium and not bitten like many people so I was actually looking into smart contract technology for a different matter and I started reading about a theorem but very

0:43interested put some personal money into it when the price was very very low that turned out to be a great investment I must admit I sold that very early just because I never had that kind of performance in such a short time so I kind of missed a lot of the upside there but eventually I ended up joining forces with some partners that when we launched

1:03a smaller crypto fun about a year and a half ago I met my co-founder and currently the CEO and CTO of into the block as Vegas and he came with the idea well first of all he came with a very interesting mixed background in both technology and finance so he was a distinguished engineer at Microsoft and then built one systems for hedge funds when that industry was the quant

1:27industry was taking off in Wall Street and in investing in blockchain technologies for a while his backgrounds finally in AI but he had this idea this thesis that became the founding thesis of into the blog which is if this asset class is really going to grow and become meaningful it needs better intelligence than what's out there because that's what what's happened with other asset

1:47class in history right additionally is the first time in history we have an asset class a financial asset class that is natively and 100% digital so that opens up the opportunity for a new form of analysis that wasn't available before in other asset classes so with that in mind we started into the blog to provide this sort of new analytics to the market

2:11and and we are that's great well thank you for the intro Alfredo and now into the block as a product you use machine learning and statistical analysis to deliver insights into cryptocurrency now have you guys built something unique in specific specific you know some secret Intel that gives you a competitive advantage over other data analytics companies in cryptocurrency yes so

2:36machine learning is a very nascent field within crypto so some of the applications that we're using are already in the product some are in the roadmap or not part of the product yet and we're using to do let's say some research efforts with with companies like coin disk with them we have collaborated in the past so one of these examples is we have a very popular

2:58indicator called in and out of the money which essentially tells you for any asset you know any address hold and it's a Bitcoin for in the case of Bitcoin whether there they would you know gain if they saw today that we lose so that's you know boring from from options lingo we call it an out of the money that one uses machine learning models to come up

3:19with cluster of addresses next prize once I haven't been privatized yet we have a very very large effort on the anonymization of the blockchain and what that means to give you a concrete example what we have done spend like four to five months building is machine learning classifier to identify exchanges on the block that will that's the the effort I was mentioning that we

3:44were we collaborated with coin design a research report for the quiddity within exchanges using the data we were seeing and also that will eventually become part of the product where we can you know take what we have today and sort of identify which players are exchanges and which player not that's great and when did you first start did you use these

4:04any algorithms when you had your crypto fund before into the block and how long is into the block been running now no so so the crypto font so there were some of the original parts of the creek the phone we became part of into the block but the fun we had close the fund actually before starting onto the block and it was a discretionary fund so we didn't use this this sort of algorithms

4:26we unfortunately have not many suits our CEO yet so so we started building the company in around November 2018 so a little over a year now we put together a team we're about 15 now mainly data scientists and engineers and we're running pretty large infrastructure for just having a year on so we have a bunch we're hosting around three terabytes of data and growing at 20 gigabytes per day

4:53so we have put together this whole infrastructure and now we we launched v1 very recently that's great and you have different types of indicators then you have at least four different types of indicators for different kinds of analytics can you talk about the different kinds that you have in the system sure so the way we have divided our indicators we have over 50

5:17indicators today and well divided those in categories which this might change in the future but if you go into the block today we have five categories the first three are mainly based on on chain data but what but it's not only just just launch and data will inject our data sources like price to come up to run this machine there any models on top of

5:37that and come up with our indicators so the five categories today are what we call financial which is that makes the amount of the money example is one of those of using blocks and data and and price data to come up with indicators we we have one section entirely just for network indicators like a trade we just launched you TXO indicators as well and

5:58others we have another one called ownership in which we basically segment the quote-unquote capital stack of a token so we say which everyone holds that token and we said meant that in two ways one by concentration so we see who are the largest holders and whether the token is distributed or centralized or not and the other ones by the time hell's always it you know a bunch of

6:20passive investors holding this acid or always the more you know maybe trading frequently and lastly we have more often indicators like exchange you know exchange data we're just recently in a couple of weeks ago we launched I got it the categories called or ebook data so basically taking order books from exchanges and coming up with analytics that we

6:41believe a pretty unique to into the block and lastly social which is completely obviously off off chain data but some of our users we're asking for so we we tried to do something original there yeah that's great and it's very interesting that users were requesting the the social and the off chain analytics and I'm curious how much importance do you believe social and an

7:06order books and an off chain data is is is there a lot of weight on that into how it may affect the future prices of the assets or not as much as on chain look I didn't to the blog we believe that an investor should look at everything right as you have different data points and make their own decisions or come up with our own strategies but the more data

7:29there's available the better now when we originally launched we sort of did not look at social too much two reasons one there were some people doing interesting things there and two we believe honestly I have to admit we believed at the time that it was you know maybe overhyped a little bit then we started listening to our customers which not only you know

7:52individual investors but also partners that are large institutions and they said look give us something on social so we we came up with again what I believe are some some interesting indicators there the Houston machine learning as well to come up with like the key topics on the news of the day and other things related to sentiment analysis for the exchange data what we do is we we're

8:17sort of combining some unchained data as well when we show you the order book so for example we have one that is market death with it which is just typical you know this is the price and these are you know different different price what are the bits and be asked on the other side but we also tell you so those price ranges this is what's being let's say

8:35offered in exchanges but also we know that on the blockchain there were this amount of addresses that were at that price so they may come and sell just to break even so sort of like combining those two pieces of those those two data sources basically to come up with something original yeah that's great and is this all surrounding Bitcoin or do you go into the depths of all of the

8:57lower market cap cryptocurrencies as well currently we're supporting about 700 assets that's we recover I think last time I checked was 13 blog chains and the tokens are run on top of them so obviously having the thean blockchain we provide data for four ETH but also for four tokens that run on top of the theorem well and how are you guys getting this information to the people

9:22that needed the most is it just direct to consumers or have you made a lot of big partnerships to distribute it widely yeah so that's a two-part question so in terms of delivery the the deliverer is basically our website which we've put a lot of effort in in the UI and the user experience because we want these data to be digestible let's say understandable

9:44by the average investor and we don't want just you know the super educated trader sophisticated I may say so so so in our website we have all these indicators in a way that is that they're easy to understand now how we get those users to come to our site we decided to partner with the top brands in crypto where those users are today and that's essentially a three buckets right

10:07exchanges data providers and and new sites so we have around 20 partners before December we had an open beta but we just closed that and so we have a freemium product and through our partners you get access to our free indicators so recently we've been rolling them out essentially one one per week since the start of the year so now you can find our data in bitcoin is

10:31crypsis late news BTC so mainly those are new sites coin gecko is already live as well and when in February we're gonna be launching in exchanges such as eToro X bitstamp and other data providers adopted for artists we have relations with all of them so we're going to be launching in Queen market cap crypto compare and and welcome gecko is already there well well

10:56that sounds like great partnerships to have and yeah we're very proud and you know our coordinates have been great and we're really looking forward to throwing all of them out that's amazing and are you using this Intel for any internal funds or you're just using it you know obviously if you have the the secret information and the secret sauce why not

11:19use it for yourselves as well sure we're getting that question a lot lately especially because we announced that probably for the end of this quarter we're going to have a chapter on predictions for price it's essentially a directional prediction not necessarily telling you exactly what the price is going to be within an hour but the quick answer is now we don't we don't have our

11:41own fund mainly because first we decided to focus on being you know hopefully the best company in the data analytics space for for crypto and building a fund requires a whole different kind of one focus into infrastructure right you need not only to raise the capital for that but also you know build strategies have risk management for fully management and

12:05back testing so it's a whole different kind of infrastructure and know-how that you have to have definitely I I completely agree and it's great to stay focused and it sounds like your team has quite the platform already and I'm guessing that you guys have quite a few amount of users are you looking at expanding more like in terms of the product you're you're constantly

12:27building in more indicators or what do users have to look forward to sure we are constantly launching new indicators and a new categories like the ones I was mentioning so every week we're so last week we launched that you take so data that I was mentioning before this week we're gonna be launching so that in an era of the money indicator has become so

12:52popular that we're doing like some we're launching related indicators to that and later this quarter gonna be launching a new chapter on futures we believe that's where there's going to be a large part of the crypto industry in 2020 and hopefully that chapter and predictions as well later on that's great and what would you say is some of the biggest challenges that you may

13:17foresee as you continue to grow into the block yeah there are a few I think we are very focused again that example that I was giving earlier on the anonymous ation of the blockchain it's been a it's been quite an undertaking we was spending again like I said like four months of the couple of a machine-learning guys completely focused on that and we want to take it to the

13:40next level and and you know not only identifying changes we're coming up with models that identify okay this is a hedge fund or an OTC desk and and that's quite difficult to do but again we're here to talk on those on those difficult tasks so so hopefully we've got it done awesome well all the best in that and if there are traders or investors that are

14:01looking to learn more information get into the subscription so they can get your indicators what's the best way for them to learn more information the best way obviously go to our website which is into the blog comm you can subscribe there for a 7-day free trial and get access to other indicators and try the platform there's also on this side at section called resources where we have a

14:25couple of videos for each one of the indicators in the platform where you can see you know how we build them and what how you can use them as a trader and lastly I would really recommend going to our medium publication our team is publishing around two to three articles a week based on data that we are seeing that is available in the platform or that we have you know other efforts

14:46internally and we and we publish them there so so it's a really great source and that's meeting the comm slash into the blog great thank you so much Alfredo I'll leave those links in the description box below that's all the time we have for the interview today but I appreciate you taking the time to be on the show and let's follow up in the near future thank you for having me

15:07Ashton and all the best of you as well you