Hello, hope you are doing well. So, a lot of people ask me if I can just go through kind of step by step here how I kind of set up my AI agentic trading pods. Uh, if you haven't seen kind of my pod theory, maybe you can go back in a different video, but uh to summarize this quick, it's basically I try to set up a lot of these smaller uh AI agentic trading strategies that kind of runs autonomous. They don't have to be like very profitable, but if you kind of stack a lot of them on top of each other, this could be like a real good way to kind of manage your trading. So today, I thought I can just go through kind of step by step how we can look for an asset to set up this trading pod on Hyperlid. So basically uh I will be using codeex today, Codeex 5.5 and we will try to set up something on Hyperlid, right? So uh basically uh the first thing we need to do of course is to find an asset or assets to trade. This could be like a pair or this could be yeah just uh bitcoin or something like that. I thought we could make it simple and just uh run bitcoin here on hyperlquid just to like bitcoin USDC just to make it easy to follow along so you can kind of implement this into your strategies. So what I want to start with like I said um find an asset to trade and that is going to be Bitcoin right Bitcoin BTC. So that is easy. So the next thing we want to do then is instruct codeex to set up the three most promising trading strategies based on the asset with uh using best practices. So I have kind of connected my um my hyperlquid here to my um API or codeex 5.5. So we can kind of control my hyperlquid profile from codeex. So I have done a video on how you set this up if you want to try the same or if you just want to follow along that's fine too. So I will be using uh GPT 5.5. I think we're just going to stay on extra high. Uh I think that's kind of good for this initial phase here now. So I'm just going to do the the prompt here now how to kind of set up these uh strategies uh based on our asset. So I don't want to keep it too complicated now. So the prompt I'm going to run is just quite straightforward. We are ready to build a new trading pod aka one isolated trading strategy on one asset. The asset we picked is Bitcoin. uh or BTCUSD. Uh uh your task is to pick the three trading strategies to test based on the asset, your expertise in trading and best practices in algo trading. Do extensive research, bring back three strategies that fits the asset class we should look further into based on your expertise. So that is quite simple the prompt I'm just going to run here. And hopefully now we can go uh and we when we move on here now we have the three threading strategies we can kind of work more on. So I'm just going to let the codeex run here now. Hopefully when we come back we have those three and then we can move further down our stepby-step list here. Okay. So we got something back here. I think it looks pretty good. So we ended up with like I said three different strategies. The first one is going to be like a volatility targeted trend breakout pod. The second one is like a intraday volatility band mean reversion pod and the third strategy is a funding premium carry pod. Uh I'm not going to go into each strategy in depth here but uh I think these are very good to start with. So yeah uh if you you can see uh recommendations start with number one trend and breakout as the primary pod but that's fine. So basically my next step now uh if we go back here uh what I want to do now is basically I want to do these two I kind of want to do it in the opposite direction down here. So we're going to collect data with codeex for these three but first I want to ask codeex to set up the best practice back testing for this trade. So the prompt I plan to use for this is basically I'm just going to say great picks. Now before we start collecting the data, please set up the best back testing setups for each strategy based on best practices from algo trading and uh adapt this to each strategy. Show your expertise in algo trading and be precise and meticulous. So this is what I'm going to run here in Codex. Now remember we are still on extra high. So hopefully now we kind of get three different setups uh with back testing. And when we kind of have those three setups, I think we're going to start collecting data we need for those back tests. And when we have that, we will of course execute the back testing. And from there, we can kind of move on or down our list, right? So, let me just uh run this now so we can set this up and we can kind of continue with our progress here. So now that we have our back testing framework uh ready. So you can see we have a couple of setups here and you can see we're going to look at trend breakout intraday and funding carry strategy. Right? So here is all the back testing we need. So basically now we're just going to start collecting the data and hyperlquid has something called historical data. I'm going to mention that in our setup here on codeex. But basically now comes the next important step. start collecting the data we need for each back test. Hyperlquid do offer historical data and I'll kind of link to that. But if we need other data points, feel free to explore and launch sub agents report back to min when you have all the data you need to run all back tests. So this step could take some time but this is of course very important because if we want like a good setup here, we need the best data we can get. Uh, I'm not saying this is going to work perfectly the first time you do it. You might even have to start setting up some uh data collections yourself or maybe you have to look other places to find the data we need for this setup. Uh, but for this uh video, I think we're just going to keep it quite simple and hopefully this is enough that we can actually run our back tests. So, I'm just going to collect the data now and I'll be right back hopefully. Okay. So you can see uh collected what is available and we started also some forward uh collection but uh basically we have all the data. You can see we have the per candles for yeah I don't know how long uh we have the USDC spot candles funding history. We looked back to 2023. You can see we have some metadata live websocket setup and everything. And we have 51 files and yeah 9 meg. You can kind of see this. We are also running a collection now. But uh for this video we probably not need that. So now we are kind of ready to look at uh the next step and that is going to be to uh start a back back testing. So now we're just going to start a back testing and see how this is looking. Just going to keep it simple here. Run the back test with the collected data for all three strategies based on algo trading's best practices. present the results in a nice structure with your best recommendation. So, we're just going to try that out here. Of course, there are room for improvement, but uh for this first uh run, I think we're just going to run it pretty straightforward and we have time to optimize and look more into it when we kind of I think we have picked the one strategy uh that looks most promising, right? So let's just run this and I'll come back when we have the results for the back test we uh are setting up here. So we have our back testing results here and we have also on report. You can see we tested all the three pods. So the trend breakout one did actually 50% net. So the sharp ratio looks yeah that's a bit high maybe 40 trades and you can see uh best strategy 4hour EMA breakout and my recommendations move forward only with the 4hour trend breakout pod not trade live yet and yeah that is basically what the results we got back from codex okay good so what I want to do now is basically move on to after the back test ask codeex to optimize and assure that the most promis ing strategy has a real edge following best practices. Okay. So this is something I like to do because uh we kind of want to doublech check our back testing just to yeah make it a bit more precise. So I uh have been testing out running some kind of prompt like this. So I'm not going to keep it too complicated. Now optimize this strategy. Aim for a sharp ratio of minimum let's say 1.2. Check for overfitting. run Monte Carlo if needed etc. Basically use the best algo trading practices to analyze and check if a strategy has a real edge after back testing bar forward-looking data. We don't really care about that now in this example. Uh run this now be precise and meticulous. Okay. So you can see we are kind of moving down the step here. Now we are kind of looking uh to see if we can actually assure that the most promising strategy we found has a real edge because sometimes uh you saw I kind of mentioned overfitting here. Sometimes the data we have with the strategy could be a bit too much aligned with our strategy. So we don't really get like a real edge if you want to call it that. But as long as we kind of refer to codeex to use the best practices in algo trading to check a strategy, it knows of course everything about this from before from its training. So we only really have to point it at that. Uh I just tried to actually mention some keywords here but I don't think it's necessary. So now you can see it's building a rigorous trend robustness optimizer. uh run walk forward stress delay Monte Carlo test optimization report and recommendation verify outputs and we're going to keep collecting some data. So again just going to let it run and let's see what we come back with here. Okay. So you can see now we kind of run this uh we run this optimization and robustness check. The result is clear reject the current strategy form. We could not optimize it anywhere near our sharp ratio of 1.2. Right? So, this is why uh I like to do this because if we just went with kind of the first uh setup uh from the first test, uh it kind of looked promising, right? It was ridiculously high, but still it looked promising. But now that we actually run this, we only ended up with a sharp ratio of 0.4. And recommendation, do not trade this strategy live, even tiny. Do not paper trade it as a candidate yet either. So yeah, basically the earlier strong result was mostly a recent window artifact. I think it was perfect that we got this now because it's really easy when we kind of take like a small sample of data. We run it and maybe it was like a smooth upwards trend on um that four window. Maybe the Bitcoin price rose like 3% or something and it looked really easy to kind of profit on this momentum trade. But now we kind of want to do go back to the drawing board with all the new contexts we have brainstorms three new more potential promising strategies we can run new back tests on work hard. So because we only had a look at the small window right uh the four window strategy that is probably not going to work over a longer period of time. So now we're going to need to go back again look for more interesting we can back test again and see if we get somewhere near that 1.2 2 uh sharp ratio we are aiming for. So again just going to go back start kind of over again and see if we can find something. So if you kind of were wondering what does sharp ratio mean? It's a measurement of an investment risk adjusted performance calculated by comparing its return to that of a risk-free asset. So if you're going to look at what does a sharp ratio of 1.2 mean? It means that the investment has earned about 1.2 two units of extra return for each unit of risk taken. In simple terms, pretty good risk adjusted performance. The sharp ratio compares uh return to above risk-free divided by volatility. And you can see a rough guide is like between one and two is good and of course everything over that is very good and that is very rare too. Right? So, uh that is kind of this sharp ratio I was talking about. So if you look back here now you can see we found uh three more strategies. Funding premium crowding reversal session gated intraday momentum and postsque squeeze volatility expansion. So again we're just going to go ahead uh good. Now rerun back tests uh on these uh collect data if needed as before. Okay. So, I'm just going to run this again. And of course, we're going to continue with the optimization to see if we can even get closer to that one plus sharp ratio. Okay. So, I'm going to let this run and let's see what we get back here. So, after a more extensive back and forward looking into different strategies, uh running different back test and stuff, uh we ended up with this one. So, this is the US late session reversal. This is a intraday strategy and the 2x cost stretch sharp of 1.12 is the cleanest near one plus uh sharp after uh friction results. We have uh the data window is short right Monte Carlo loss probability is still uh meaningful uh last for walk forward window lost money. This is not a production ready, but this is the best we found in this video. But basically, uh for the video sake, uh we could have gone further down looking for more, but uh I just want to create this pod now because uh this is something I want to do. And we're going to fire it up and maybe build a simple UI so we can track this. So, I'm just going to go ahead with my prompt here. Just going to keep it simple. We are ready to set up this strategy live for a better and real data. Build out the script needed to run this as per the strategy. Pick the best code language to run this in. Also create a single HTML UI that tracks this trade. Simple dark mode with an HTML terminal style. Look, execute this. Now, I'm going to keep it really simple just for the video sake. So, basically, I'm just going to repeat what I said in my previous video. This is just going to be one small pod that is going to run in parallel to a bunch of other similar strategies like this. So the goal is that let's say we have 10 different pods. If seven of them are profitable and the other three are not, uh we can of course still make some money, right? And as long as we stack these pods, each pod does not really have to be have like a insane return as long as is it is profitable. And the other good thing about this is that you don't get really hyperfocused on just one trade. you spread it out and you kind of forget uh it's already there. You just come back to check the numbers and to confirm that it's still profitable. If it kind of goes into the red over time, just swap it out with a new one you find. So, this is what I'm going to doing uh the next uh the la the last part of 2026. I'm going to keep building out this and we're going to track them over time to see how they're doing. I'm going to do an update video on my pod setup soon. But uh let's just build this out and let's get it up and running and actually see if we uh are live with this and uh we can get this pod running for some 24 hours or something. So now you can see we have our uh session or pod running live here. You can see we updated some UI here. We have like you can see my balance here $690. Uh we have the signal the market we are in. We have some events here. You can see live client ready. And yeah, but this is like a this is not in session, but that's fine. Uh, but I'm going to keep running this for a while just to see what happens. But like I said, I don't have any big big um expectations for this setup, but I'm going to keep it running for a while. But, uh, I want to do a test. So, here you can see I am on Hyperlid. This is kind of my portfolio. You can see my balance 689. So, uh just do a quick um test trade live. Um so, just buy some a Bitcoin position and cancel a position after 10 seconds because I just want to see that we are up and running. So, everything is working. So, we should be able to see our position pop up here. And from there, we kind of know that we are all wired up and everything is aligned. And you can see now we entered a position here. $14 on BTC. Remember, we should cancel this autonomously after 10 seconds. So, let's see if it goes away. Yeah, perfect. So, now we know that we are online and we are ready to kind of identically trade on Hyperlid with the setup we have. So, I'm just going to do a final thing. So, just explain the pod strategy again in simple terms so we all can understand it. So, we can see this simple example here. Bitcoin drops hard from this time frame much more than normal. The pod says uh that move may be stretched by Bitcoin hold for about 2 hours then closes. That is basically the strategy and looking at the data we pulled. This looked promising and you can see the setup is uh going to be Bitcoin only 15-minute candles. Uh we're going to do a $50 entry with no leverage and we're going to exit autonomously after uh the fixed uh position hold period. Right. So that is the pod and it is up and running. So I'm going just going to keep it for at least 24 hours, maybe a couple of couple of days just to see how it goes in that kind of timerestricted window. So it doesn't really matter if it it's very boring, right? You don't really want to pay attention to this because it's just a small window each day. But as long as it's running in parallel in the background, it doesn't really matter if it can make us like a small profit. Even like $5 a day is profit, right? So, we're just going to keep this and I will update you how this goes now. So, basically, this uh was the strategy uh or was what I wanted to talk about today. how I kind of set up uh one pod uh on Hyperlid that I'm just going to keep running in parallel. Of course, I just picked like a simple asset today. Uh but we could have made this more um complex. We could have maybe like looked at two different assets. Maybe we looked for some correlation between I don't know Brent oil and maybe I don't know something else Nvidia or something like that just to see if we can find anything. But that's going to be in future videos. If you are interested in that go check it out. Uh also I want to mention we are running a discord. So if you want to come by talk about aentic AI trading, AI automation, how we can actually use this to make some uh income side hustle, call it what you want. Drop by the Discord. Hope this was enjoyable and that you maybe learned something. Maybe this gave you some ideas. I'm not saying you should copy me exactly, but maybe this gave you some ideas of what you can do in your uh yeah strategy. So, thank you for tuning in and I will report back on how this goes in a future video. So, give it a like, subscribe if you want to see more and I'll see you again soon. And by the way, go Norway. We are playing soon in the World Cup. So, yeah. My