Failed twice, I was super close to something but nothing ever materialised [I will not promote]

Hey guys I am Div (Divyesh), I’ve been a founder twice while I was in university, I am now back in the UK looking to start a startup or join one or find people who are interesting also trying to start a business. I got really close to materialising something but failed here are some things I learnt, this is kind of like a rant but maybe you will learn something from me. So I really got into entrepreneurship back in 2024, I think was 19 or 20 at that time, I was part of the Microsoft Student Ambassadors programme where I came across the Microsoft imagine competition, when the competition came out, I was in India during that time for my University where there was this issue of Buildings being torn down despite being given approval, this was due to the buildings being classified as unsafe later on or human error, So I had this idea of what if there was a deep learning model that could read architectural blue prints and find errors in compliance and later on, I thought given a prompt (or JSON) what if we could generate architectural blueprints. That’s what I did, in hindsight my solution was that you could convert some CAD formats into ezdxf and vice versa and then you can train a transformer on it, I tried to get into Graph based and Edge based Deep learning models but I didn’t have the time nor knowledge to spend on that, but given this idea and the MVP I made I managed to get the highest tier of azure funding credits through the Microsoft Founders Programme ($150000) where I had access to H200 GPUs I was training I was setting infrastructure I was in talks with potential clients! But nothing ever really went through! They couldn’t trust me or my product I wasn’t able to convince them with my pitch maybe, the fact that I was still a kid from university could be a reason, the fact that I had no “Founder Lore” could be a reason as well. In essence, I had a broken distribution channel, I had no GTM strategy and my assumption was that, given the type of product I was building it was hard to trust someone unlicensed with it. So I expanded, I tried to get a cofounder this time, I got onto F6S, I tried to recruit from LinkedIn, I tried to find cofounders from the YC connect platform, and this time my strategy was since I have so much compute available I can offer ML and AI model solutions to companies, models that train on their data and then I can provide subsidised inference for a time period, I probably started doing Forward deployed engineering before it was coined. Now this was when I got the closest to locking in some clients who were ready to pay, but as the saying goes “anything that can go wrong will go wrong” my startup was not registered, in India registration is a really long process which takes up to 3 months, I knew that this would come back to bite me and as a uni student I didn’t want to register and go through that massive process and file annual tax reports etc. I thought I could maybe register as a soleowner later on and bag a deal, I lost 2 deals this way, because some firms were not ready to sign a contract with no backing, when I started my registration I could not find a client again, for 6 months I had a tax registration for my firm, my parents didn’t know that I spent all the money I had saved from internships on my project and I ran out cash. I shut down, I then started something called sociobot this time I thought GTM would be easy, and no SaaS products are even harder to get people to adopt, B2C is super hard to prove good traction and by this point I was exhausted. Here are the lessons I learned: Always have a cofounder, being a solo founder may work but trust me you’ll lose motivation fast you’ll feel really sad after a point and you’ll be burnt out as well, don’t bear the burden alone, I was in such a bad spot at one point I had to regain my sanity by talking to strangers and trying make random friends since I lost many friends during my entrepreneurship phase Sell the product before you make it, (some people may disagree) but your first occupation as a founder needs to be a salesperson before a product specialist, you need to convince people to want your product and you need to be able identify the products shape from constant discussions with potential customers before you even start. Be a go getter, everyone complains about not having capital to start, this is false (to some extent), there are so many programmes out there that give you a really good starting point, you just have to look for them. Don't over generalise your product, dont try to address the entire TAM, because you are going to just confuse everyone who uses your product, don't try to add features to make it a full solution for every single person, focus on the niche and build a really focused solution even if it means compromising on some part of the market TL;DR : I was a founder during university, I was really close to materialising a successful startup and because I was immature and despite having a really good chance I really effed things up with bad decisions and things went down the drain. I started again and failed and I’m now in the UK trying to start again or maybe be a part of someone else’s journey.