How to pass the project accelerator and what it gives
Recently we tried to get into the accelerator with our project. We created a very simple solution a long time ago: “Development of a domestic Smart Home and Internet of Things platform based on the ESP32 board, as well as creation of a corresponding educational course”
If anyone is interested, here is the latest information.
The NTI became interested in the project, I heard firsthand what it gives: selected projects that get good scores on the demo day will receive a subsidy from the funds that finance the NPO organizers. We were invited to the accelerator as part of the Archipelago program.
If we had previously gone through pre-acceleration with another project in 2 weeks, and I know from hearsay that they can prepare for months in an accelerator, then here the deadlines were super tight. When I received an invitation by mail, I realized that there was already about a week and a half left. I started by assembling a team, because it was a mess. I naively assumed that in such a short time we would have time to gather everyone who was interested in any information on the project page. While I was collecting contacts and writing to everyone, I was simultaneously attracting a team of people I already knew, with whom we had worked well together.
1 Error
Don't think you have plenty of time. If the hackathon gives 2-3 days, and you've recently gotten used to that time, remember that 2 weeks go by faster because the team members are working in their main places.
I barely managed to gather the guys to defend us, who blindly believed that we still had a little time left.
The first days are when you think you can do everything
We had assembled a core of 3 people, but I saw that there were 10 more candidates on the team. We started discussing the first steps of the project. But we couldn't get everyone together. I naively believed that I could afford to personally explain the essence of the project to everyone. But I realized that I had wasted all the initial energy on this.
A few days later I realized that we were short of people for AI, so I brought in two developers. I had to tell them everything again. Of course, half of the 5 people decided not to go to the calls with the curator, thinking that there was still plenty of time, and the rest would decide what to do without them. I spent another 50% of my engineering energy on these guys, just calling them when they didn't pick up the phone.
Overall, we finally got together for the 3rd call with the accelerator mentor, but I kept hearing from someone that he couldn’t, and from someone else that he was interested and would be happy to join our team.
We suggested that we sell not 1 set, but 4, to smoothly begin immersing students in the Internet of Things, and finish with datasets for training models. For example, to assess the quality of soil in growboxes and greenhouses, to assess crop yield indicators.
They also brought in some projects from the past, well, it’s logical that they fit perfectly into the concept.
2 error
Thinking that you can explain the same thing to everyone, and there is no need to hold a group call every day, it is enough to decide on something big.
You can everything
The thing is that, according to the plan, the accelerator probably thinks that you are already working on the project with a team and all participants are actively working. In reality, this was not at all the case: it was our project that was developing from the angle of artificial intelligence (Microservices with Docker containers and neural networks, victories in hackathons). But the electronics part was done a long time ago in the form that was interesting to us. Then – just look for an investor and pump everything up.
During calls with the mentor, I began to hear that she was dragging our project out without offering funding. She was trying to push the project as much as possible, devaluing what we believed in. And to push it into a new phase – from toys and robotic crafts to a platform for managing engineering projects.
In general, my vision was already like this, but they really helped us think about many things: who is our client, who should we make the first 3 sales to….
By the way, the first thing they ask you about is what you personally want from the accelerator. As I understand it, this is a mass phenomenon, and if you do not propose a question about the first test sales, as we proposed, then they will write something like “prepare a project and presentation for possible communication with investors” in the acceleration goals.
Our team started with dreams about the day being a meeting with investors, when you are offered your first money, or they tell you what is still missing to finally give you investments.
By the way, my personal achievements are that I am a scholarship holder of both Eranet Mundus and Erasmus. I am a holder of a state award, numerous letters of gratitude with monetary incentives, I worked on the Umnik grant and applied to Start. It is logical, what goals could be for me now: to win the start, finally, at the end of my still young life, to go through a cool accelerator, to receive investor money and issue my shares. As if I were winning in cooler things, it is high time. Yes, the article is on behalf of Evgeny Nikolaevich Rychkov.
What awaited us next? We came up with a large scheme, how the Internet of Things server iot.tfeya.ru stands at the heart of our entire project, surrounded by both a project base and a system for taking over these projects. Projects should grow and become more attractive to new engineers.
What mentoring gave
I really wouldn’t think that it’s worth selling not the takeover of projects for 5,000 rubles from engineer to engineer, but rather that you need to sell startups that have grown into systems, which are often difficult for the founders to continue, but which can cost half a million rubles or 5 million rubles on the market.
The mentor picked out some cool words, helped me remember terms like pet projects and impact projects, and reminded me about information about generations of people: millennials, z-generation.
We have put together a beautiful value proposition:
We help IT engineers and developers (type Y) in the field of the Internet of Things (situation Z) create innovative solutions and interact with like-minded people (problem P) through a platform for collaboration, knowledge sharing and obtaining project funding (technology S) and get opportunities for professional, personal and material growth (value V).
We also decided something like this.
The problem of low competence of engineers at the intersection of electronics and artificial intelligence is becoming increasingly relevant. If there are courses in the niches themselves, there are professions of radio engineer, ML engineer, electronics engineer, then it is obvious that in the modern world with the development of artificial intelligence, more and more tasks appear that can change a person's life, add good habits and rules and reduce laziness and an unhealthy lifestyle.
There should be projects, due to which not companies, but any individual person will be able to earn money, create food and benefits. And which will not just turn into something accessible in 20 years, but will be able to appear quickly and be implemented not only by motivated specialists, but also by anyone who is interested in the gamified development process. Alice should be smart and have the functions of a fitness trainer, an advisor for improving life and daily routine. Her decisions should not be based only on the experience of the transformer, but a component with a RAG model is needed for additional training on the data of a specific user of the system. Solutions at the junction of electronics and AI can significantly change our lives not only in a smart home, but also in the country and in other places of stay.
The creation of relatively simple solutions (smart lamp, sensor data collection station, smart socket, cooking automation devices) based on ESP32 has become a popular direction with the advent of devices like Alice. With the advent of voice control, you can “communicate” with home appliances as objects or even “residents” of your home. For example, for many, the sequence of commands “Alice, turn on the vacuum cleaner-kettle-coffee machine-TV-light in the kitchen” is already common.
With the development of the Smart Home, it is also important to be able to plan when certain devices are turned on and under what conditions. We can immediately plan when the house will be cleaned, when we have breakfast, when we need to turn on the TV and take a break from remote work and watch a movie while drinking tea. For such a complex interaction of devices, with maximum distraction of a person from these processes, in addition to direct supervision of the state of devices, the Internet of Things is becoming more and more deeply embedded in our lives. Many devices do routine work, allowing human time to be spent on solving larger-scale problems.
We are currently working on a course based on the ESP32 microcontroller, which will teach the user to assemble their first project – a smart box equipped with sensors for assessing the health level of the room. But simply teaching how to connect your device to the smart home system is our minimum task. We are focused on creating devices that will create datasets with which you can train neural network models to make certain decisions.
We believe that everything here should start with small and extremely simple, but working functionality created by random people. The support team should be a cut above in experience to control the process. For example, one developer who accidentally ended up on our hub is engaged in assessing the condition and fertility of the soil. Another determines the yield based on such input parameters as illumination, humidity, temperature, type of seeds, variety, and so on. Not only electronic projects and prototypes are being improved, which are in some ways worse than industrial samples bought in the store, and in some ways superior to them. Algorithms are also being improved, which are becoming more and more accessible to a wide range of engineers who adopt projects and improve them. Hub users can exchange developments using API interfaces, the first datasets will be expensive, but after the first sales they will become more and more accessible so as not to pay a subscription fee for another microservice solution, but to adopt it for yourself.
By analogy with relatively less useful services that solve non-technical issues of humanity, such as RuTube, YouTube, OneRPM, and also by analogy with services such as Apple Store, RuStore, we are moving towards a solution with monetization, where anyone can earn money by developing their project and participating in the rating of solutions.
We have a specific goal to complete the course within the system of 4 courses for quick immersion in the Internet of Things and neural network solutions, launch the sale of the course and the kit for it. As part of the project, we are making a corresponding prototype of the device and are going to bring it to market. This is a smart greenhouse, similar to the Wega project, but developed within the hub and according to its rules not by our group working on the project completely, but by adopting our MVP, which will be completed, by other random developers.
One of our key differences from what already exists is the emphasis on small, working modules that can be purchased on the platform and from which you can earn money without thinking about where to find a customer. These should not be fully working projects, in which there cannot be a human factor of the type “I gave all the data, copy the project, and I wash my hands of it.” Our goal is to make an extremely simple and understandable system for adopting and developing projects that are not available in OpenSource. We gather like-minded people and resources so that for random developers coming to the hub, the process of working on the next project is understandable, defined in terms of goals and financial components.
A little bit about our strengths. Historically, we were almost the first to test this business in Russia (possibly in Europe as well, one thing is certain – we were hardly the first in China), the roots go back to 2015. There was a Blynk solution, there were no IOT servers yet, within the Technofey developer community we were looking for orders and fulfilling obligations in the IOT field. One of the first projects that Evgeny Nikolaevich Rychkov started with was a smart hotel on Lake Baikal, which was developed on the basis of the ESP8266 board and Blynk. Previously, a smart home was more accessible to the rich, and solutions and modules were extremely expensive, although the owner of a cottage could use it. Basic functionality was provided: lighting control, floor heating, contactless switching on of devices and appliances.
It took several years since then before relatively adequate resources such as Iocontrol began to appear. They offer developers to connect a board with a WiFi module to the portal and monitor the values received from sensors connected to this board. In 2018, there was a “hype” for IoT in Russia, a conference of the same name held at VDNKh is known. After 2020, smart speakers and sensors from large Russian companies and startups began to appear on the Russian market, for example, Yandex Station, a temperature and humidity sensor, a smart light switch. Videos from companies, in particular from Yandex, appeared, showing an apartment with smart gadgets and devices that automate a person's life. Where you can pour coffee by simply asking Alice about it. Of course, this has roots in the corresponding American projects, in particular, you can familiarize yourself with the history and capabilities of Alexa. Only in 2022+ did IoT services from large companies (Sberbank, Yandex, MTS) begin to emerge.
Our advantage is that we are maximally focused on popularizing work in the direction of IoT in Russia, we have almost proprietary experience in order to dictate new ideas for the Russian market and are at an interesting intersection of electronics and IoT, industrial production and the education system. We have experience working with industrial production and cooperation with companies, but our goal is to help a specific engineer with earnings and monetization of his activities, and not companies. The project is a unique solution both in terms of the idea and in the energy that we give to developers and those interested in the project. One could say that we and our ideas should be completely absorbed by companies such as MTS, Yandex, Sber, and only time will tell.
We need more workers, and we don't mind if within the framework of this same project you work on other devices in parallel and together with us, or develop what we need now, together with us. The main idea is that we are forming an ecosystem of courses and solutions, and in the vision of the project we want these devices to be really more interesting to buy from us than on the market, so that they are somewhat more competitive and solve non-standard problems that modern smart home devices on the market do not solve. We want this to really help a person make his everyday life easier and save his time at home.
Overall, I wouldn't say there's much that's new here. Or even that the terms or words are new. Everything was there before. It's just that the mentor immediately told us what and where to add, what was missing, and what to remove.
We noticed things we wouldn't have noticed without mentoring. You can write news in the project. You can show what has been done and what still needs to be accomplished.
Weekend
I gave the electronics engineer the task of drawing a PCB for a waterer – a smart watering device for an automated greenhouse, which we had developed on our knee a long time ago. For sale, we just needed a normal board, not the flawed one that we had before.
I gave the back-end developer the task of running a sketch of a human on AI on the back-end with RAG+Mistral. I gave Sveta the task of finishing the presentation. I also loaded the back-end developer with a roadmap. I started thinking about it myself, calling everyone and asking how things were going and whether they understood what I saw.
Disappointment
The weekend flew by, the presentation seems to be there, but something is not quite what I wanted. The backend seemed to have launched everything, but when I decided to take control and ask if there were any errors when importing from Nikita's sketch, I heard that there were none. Under the fact that ML should go to the server, bitch, the backend assumed that he had registered greps npm installations there, and now it's up to the neural networker, who should import and launch, bitch, his ipynb himself! Of course, I call them all three, but there is a conflict like shifting roles.
Okay, I ask the backend developer, what's the outcome with the roadmap? The answer is something like “Zhenek, go ahead and write it down, there's some important info coming up.” Of course, I was a little timid, then I reached for the phone and started asking “maybe I should clarify what exactly should be in the roadmap rectangles and what color should they be when you call me? And you should write it down”?
In general, the continuation was very funny. Everyone took responsibility. But not the one I proposed, but the one that each person's consciousness accepted crookedly. Someone did not take this responsibility at all and tormented me with requests to call again and first better understand what we, bitch, want. Although on the group call, where these people successfully did not connect after 2-3 invitations… In general, I did not understand at all where the team was, where I was behaving incorrectly. After all, often everything is not so terrible. Everyone is so smart, everyone has work, no one can clearly say what to rely on him for.
The team is gathering
Rag+Mistral, with my efforts, we finally brought it to the stage of not despairing. They sent me the PCB, there was even a slot for connecting a submersible pump. We noted the projects, came to the conclusion that we, after all (hurray!) do not make toys for children, but make a system that gives birth to projects in the field of IoT for business. I was no longer in a hurry to change the composition, because there were a lot of people, but there were about 5 days left before the defense.
In the end I got the return, the presentation was ready on the last day, but there was also a roadmap. We used Anton's recommendation from my team Down To Up in market research, but I was surprised that the mentor helped us a lot with the market analysis. SOM, however, we did not consider.
Summary
In the end, it is impossible to tell anything at all in a 3-minute pitch. In the case of our project, it seems to be packed, but we only realized at the very last moment how to describe it in 3 minutes. But those who are not familiar with the project may call it either a genius move or the ravings of a madman. I wrote down a slightly more extended pitch here:
https://www.youtube.com/watch?v=GyyHPCqW21Y
We successfully defended the projects during the demo day. It seems like they even showed interest in the project, but they didn't offer us anything right away. They just continue to communicate and talk about their interest in further development of our project. Well, what can you say, you think that now they will get the first money, and they will lead us to the pitch to investors.
In practice, it's not that there's not enough time, but the accelerator itself, which is held in a short time, gives you a very good vision of what and how to do for the project. But it doesn't seem to reach investors, not only in our case. Now we're sitting and looking for investors, there is an information cloud for the project, which is available for the masses to review, for which, of course, thanks to the mentor.
I would still recommend participating in accelerators for people, because it does, however, bring together what you already have and moves you forward a little, gives you a boost.
