Subscribe to AIMinds Newsletter
Stay up-to-date with the latest AI Apps and cutting-edge AI news.
Subscribe
Share this guide

About this episode

“We're using Deepgram for the speech to text API. It has saved so much time for the SLA that we have for our customers, I mean, the turnout time of their clips. So Deepgram converts an audio file of 1 hour long into a text in just two minutes.”

— Sheida Mirjahani

Sheida Mirjahani is the COO and Co-Founder of Flowjin, which is one of the top AI-assisted video editors for content creators and marketing teams. She co-founded Flowjin 18 months ago and before Flowjin, she was a content marketer, analyst, market researcher, and business developer and helped +10 aspiring entrepreneurs to launch their new venture and raise +$2M in less than a year.

Listen to the episode on Spotify, Apple Podcast, Podcast addicts, Castbox. You can also watch this episode on YouTube.

In this AI Minds episode, Demetrios is joined by Sheida Mirjahani, the COO and Co-Founder of Flowjin. Sheida recounted her journey from Tehran, Iran, to her current ventures in Toronto, Canada. Her transition into tech and entrepreneurship is one of the highlights of her story and she shared valuable lessons.

She did not only share her background but also discussed Flowjin's latest upgrade, Flowjin 2.0. This episode showcased how Flowjin is transforming content creation using AI-driven solutions, particularly in audio and video mediums. Sheida also addressed the challenges and successes of integrating AI into their technology.

Flowjin's story highlights the importance of innovation and adaptability, with their collaboration with Deepgram playing a significant role. She explained how Deepgram's startup program has boosted Flowjin's efficiency and capabilities, contributing to their success.

Fun Fact: Sheida Mirjahani and her sister co-founded Flowjin after realizing the need to create a startup of their own, as they felt like they were "giving birth to a child" with their previous freelance projects and wanted to keep a project for themselves.

Show Notes:

00:00 Moved to Toronto for a new startup.
06:04 Launched agency, helped 12 founders, raised $2M.
08:59 Switched from audio app like Patreon.
12:12 Tough decision to delete code, follow influencers.
16:44 Turning podcasts into viral social media clips.
20:22 Switching to Flowjin 2.0, built AI infrastructure.
21:37 AI market changing fast, poses big challenges.
26:44 Deepgram's speech to text API is a time-saver!

More Quotes from Sheida:

“It was really challenging because imagine coding an application for a year and doing marketing and acquiring your first batch of customers, but then you decide to let it go and delete all the codes. So, it was very challenging, but it was the right decision.”

— Sheida Mirjahani on starting a tech startup

“This market is very rapidly changing. I mean, you've seen that in 2023, there are a lot of AI startups coming to the market, and also the AI infrastructure is growing insanely fast.”

— Sheida Mirjahani on the growing AI industry

“We help people who want to create audio content, specifically coaches, freelancers, the ones that have an audience on a social channel, but wanted to try a different way to engage their audience and make money out of creating content and selling subscriptions, to their knowledge, in audio.”

— Sheida Mirjahani on using Deepgram

Transcript:

Demetrios:

Welcome to the AI Minds podcast. This is a podcast where we explore the companies of tomorrow built AI. First, I'm your host, Demetrios, and this episode is brought to you by Deepgram, the number one speech to text and text to speech API on the Internet, trusted by the world's top startups, enterprises, and conversational AI leaders like Spotify, Twilio, NASA, and Citibank. In this episode, we are joined by the COO and co founder of Flowjin, Sheida Mirjahani. Welcome, Sheida. How you doing?

Sheida Mirjahani:

Good, Demetrios, how are you?

Demetrios:

I'm great. So where are you calling in from? I feel like it is on the same side of the world as me.

Sheida Mirjahani:

I'm guessing so. So I'm calling from Toronto.

Demetrios:

Oh, all right. Nope, I was way off. I'm like a couple thousand miles away from you. I am in Europe, but that's awesome to hear. I want to dig into your story of how you got into tech and started creating companies and all of that fun stuff. So maybe we could start at the beginning.

Sheida Mirjahani:

Sure. So I think my story of getting into tech goes back to when I was a 19 year old girl. So I was at college and I was basically studying my bachelor. And one day my sister came to home and said that, what are you doing with your life? Why are you not having a side project? She's actually my co founder.

Demetrios:

Oh, that is so great. Yeah, I want to hear that story next, but, yeah, sorry for getting you off.

Sheida Mirjahani:

Sure, yeah. So one day she came home and said, okay, I know that you are studying, but I think that if you work on the side, it would be best to start as soon as possible to get to know what you want to do with your life and so on. So I went on Twitter and I just saw that there's a job post on Twitter and there was a startup at an accelerator back in Tehran. So I'm originally from Tehran, Iran. So this is where I applied for my first job in tech, which was a crowdfunding platform in. So that was when I first started to work at a tech company. It was a very good experience. And since then I have either have been working in startups or with startups or on a startup right now.

Sheida Mirjahani:

Yeah. So here I am.

Demetrios:

And how did the move, how did you end up in Toronto? What's that story?

Sheida Mirjahani:

I moved to Toronto two years ago, and the reason that I moved is that I wanted to have a little bit of different experience in my career. So I wanted to start a startup with my co founders and we wanted to accept a challenge to do it in North America. So we said that, okay, let's apply for a startup visa program, which back then was a popular program in Canada. And we got accepted into an accelerator which is located in Vancouver. And then I moved here two years ago and since then I moved to mean I didn't move to Toronto, but I landed here. I never went to Vancouver because the tech scene in Toronto is, in my opinion, is a little bit better than Vancouver. I mean, there are a lot of facilities here, a lot of incubators, and the ecosystem is a little bit mature comparing to Vancouver. Yeah.

Sheida Mirjahani:

So here I am in Canada. Toronto to Canada. Yeah.

Demetrios:

There's some great universities there too, and I know that they're working on all kinds of cool stuff when it comes to AI and computer chips, all of that fun stuff. And the meetup scene is great there. The diversity of cultures you get in Toronto. I enjoy the city thoroughly. I think it's a great place. I've actually been to a few conferences there and loved it every time I was there. So I can see why you ended up just hanging out there and enjoying it now. Can you talk to me about some of these startups that you were part of and what you learned over the years while you're working at different startups before starting? Flowjit, of course.

Sheida Mirjahani:

Yeah. So for a couple of years after my college, I was working startups. They were in different industry. Like, one of them was a crowdfunding platform, the other one was something similar to booking.com, sorry, a similar platform to booking.com, but the local one in Iran. So there were so many startups that I either worked at them on a project level or like a full time. And I actually tried so many fields. So I was a content manager, I was a product designer, I was a data analyst in one of them. So I was exploring different aspects of building a business, but I didn't know back then that I'm doing this.

Sheida Mirjahani:

I was just following my interests, my passions. After a couple of years of working in startups, my sister and I decided to start a very small bodic agency of our own. So back then we were doing some freelance projects to help people who wanted to become an entrepreneur. And they were an expert in their field, but they didn't know how to start a startup from scratch. So we were helping them with market research and also we co designed the MVP of the product together with them. And also we had a freelance team of technical people who could code the first version of the product. So we helped around like twelve new founders who wanted to start something. And then they have also, in one or two years, I think they raised more than $2 million after doing the project with us and running the company on their own.

Sheida Mirjahani:

So that was a very happy moment for us and incredible. Back in 2019, we decided to start Flowjin. So I'm going to stop right here if you want to ask me any questions.

Demetrios:

No, I really like how this easily flowed from one to the other. And I am going to let you keep going. I want to hear this story. You left me on a cliffhanger.

Sheida Mirjahani:

Okay, cool. So I think it was COVID time that we decided, like me and my sister, we said that, okay, we are helping these entrepreneurs. And the way that we worked was that we designed a concept, made it into a product, and then helped the founder to code the first version and then go from there. So we always felt that, okay, it's like we are giving birth to a child, and then we're giving away. And it was feeling that, okay, we want to have something that we don't have to let it go. So we want to start a startup on our own. So at the start of COVID I think it was the time that we decided to start something of our own, which was Flowjin. Actually, it was Flowjin version one, and now we are at Flowjin version two, which I'm going to explain later.

Sheida Mirjahani:

But back then, we created an application that was kind of like, I would say Patreon for audio creators. So we helped people who wanted to create an audio content, specifically coaches, freelancers, the ones that have an audience on a social channel, but wanted to try a different way to engage their audience and make money out of creating content and selling subscriptions, to their knowledge, in audio. So we worked on the application for over a year, year and a half, but we decided to let it go and to pivot at some point, which was in August 2022, where the new flow gene was born, which is the flow gene version two. So this is a whole new idea. So the reason that we pivoted is that we understood that the real pain point of content creators is not actually creating audio content. I mean, it's not creating content easier because the application helped them to record the thing. Very easy, just like sending a voice message to your friend on WhatsApp or any other messenger gaps, and just put it on a playlist and sell a subscription to it. So we thought that, okay, after talking to so many of our users, they told us that they're not looking for a better way or an easier way to record things or create content, they are already doing that.

Sheida Mirjahani:

The problem is that after a content is created, people have hard, I mean, these content creators have hard time to build an audience or build a listenership for their podcast or to build know followership for their YouTube channel, you name it. So that's why we landed on Flowjin version two, which is a platform that helps you create content for your socials very easily. I want to stop there so that if you want to jump in or.

Demetrios:

So many questions. Yeah, I really like that you broke it down for me and it feels like one thing you were doing from the start was going out there and interviewing the people that are using your tools and seeing, is this tool valuable? Is it doing what we think? Are people using it how we think they should be using it? And if so, what are some things that they want? Can you break down that process and what it was like going out there and connecting with your users and how you managed to. I'm sure it was a big decision that you made when you decided to go from Flowjin 1.0 to 2.0. And so what were some of those conversations like? Also.

Sheida Mirjahani:

It was really challenging because imagine coding an application for a year and doing marketing and acquiring your first batch of customers, but then you decide to let it go and delete all the codes. So it was very challenging, but it was the right decision. The process that we talked to the customers was that basically back then we were following a lot of content creators on Twitter. And actually the idea of flow gene came from following all these people and seeing the new trends and coming up with the flow gene 1.0 idea. And then when the first batch of customers installed our application and started to using it, we tried to message all of them one by one to get them to do a call with us to understand, okay, how is the experience for them using Flowjin? Are they enjoying it? What are the pitfalls? Back then we understood that they're saying that we have a challenge, first of all, bringing our audience, for example, from Twitter, from YouTube, to a new app that nobody knows about. And we have to convince our audience to come here, listen to a new content. And also, they were not really comfortable selling their content in a very direct way. Like when you are a YouTuber, you are earning money passively, not from your audience.

Sheida Mirjahani:

Like your audience isn't paying you anything, you're just earning from ads. But when you want to do a patreon or using Flowjin 1.0 they had to convince their audience to pay them in order to listen to their content. So those were the first category of challenges. But the main thing was that this was not actually a need, that they wanted to continue using the app. So the app wasn't actually solving any need for them. And then we understood that, okay, these people, they are creating content on Twitter, on YouTube, they're doing a podcast, but it takes so much time for them to promote their content on socials, to gain awareness of what they're doing, and to just spread the word. So that's how we pivoted into Flowjin.

Demetrios:

2.0, and presumably all of that market research and the studying of what you learned beforehand when you were giving the babies away to the other founders. As you mentioned, that was something that you were doing before you started Flowjin 1.0. And then you realized, okay, we need to continue to make this product better. And you continued to reach out to people and ask them. And you saw that the highest value potentially isn't trying to get these creators paid by having them bring their audience to a new platform. It is helping them in the platforms they're already on. And I suspect that a lot of these creators have a large audience on maybe one of the platforms. And what your big value prop is now, if I'm understanding this correctly, is you're saying we can help you grow your audience on the other platforms where you're trying to grow it.

Demetrios:

Also, because no creator wants to be beholden to just one platform. And I think everyone knows that. It's good to diversify, whether that is starting a podcast or doing more videos. If you're currently only posting on Instagram, that's probably a bad sign. Maybe it was okay in 2016 and 2018, but now all of the influencer revenue that you were getting, it may have dried up if you haven't diversified across different platforms. So I see that value prop being much stronger, and it makes complete sense on why you would pivot in that regard. Now, what exactly is it doing? How is it making lives easier?

Sheida Mirjahani:

So, you are a podcaster. I'm sure that you are trying different ways to promote your podcast after it's published. I'm assuming that the majority of the work that you do is after recording a podcast, which is editing the podcast. Either you do it on your own, or you outsource it to a freelancer, to a team, and then it gets published, and then what? You want to get listeners, right. So the way that we help you, as an example, as a podcaster, is that we turn long audio or video recordings into short viral clips and it's optimized for any social media platform. So the way that it works is that you put in your YouTube link, your podcast link onto our platform, and within 15 minutes you will get multiple recommended clips generated by Flowjin automatically. And then by selecting a couple of set, sorry. By selecting and editing a couple of text, you can customize the clip and make it perfect for the social channel that you want to post that.

Sheida Mirjahani:

So that's how flowchain 2.0 works.

Demetrios:

Oh, that's awesome. And is it adding different visuals to it, or is it taking the clip and just making sure that the clip is the best clip? Like the Golden 15 20 seconds from the long form audio.

Sheida Mirjahani:

So we both support audio and video podcast. So it depends if you are creating an audio podcast and you're using Flowjin. We are creating sound visuals, like creating an avatar for a speaker, because we want a visual for the podcast. Right. After analyzing your content, we detect speakers. So we prompt you to tell us who are the speakers and then we create an avatar for them. And the visual looks like showing the active speaker on the clip whenever they're talking and it goes back from the active speakers. This is for the audiograms, but for the videograms.

Sheida Mirjahani:

For now we are just adding captions that you can customize the styles, but later on we will add brolls and other visuals that are right on the context of the clip. Because we understand the content, we can also add visuals to the clip as well.

Demetrios:

Very cool. Yes. This is something that me as a podcaster I'm doing already. And so if there are ways that I could get technology to help me do this, it would be super helpful. And maybe one thing that we could do with this podcast right now is we can run it through Flowjin after it's already, and then we will add what that looks like onto the YouTube channel and we'll put a link in the description below also.

Sheida Mirjahani:

Yeah, I'm very excited to see how your clips turn out, especially this episode that we're doing together, because I haven't seen myself in a clip.

Demetrios:

Nice. Well, this is going to be a perfect. Yeah, this is very cool. So now talk to me about some of the challenges that you've had when it comes to building with AI.

Sheida Mirjahani:

Sure. So you might imagine that pivoting from Flowj 1.0 to 2.0, we had to develop some AI infrastructure to be able to do that before then. In 1.0 we didn't use any AI. So the first challenge was that we were a team of co founders and none of us are like AI specialists or machine learning experts. So the first challenge was to create a proof of concept as fast as possible with the things that we had. This was before Chad GBD time. So the first challenge was to stick together what we could find to do as a proof of concept. But when we created the proof of concept and then just talked to some customer, to some potential customers, and they paid us, you know, the first few bucks, the first challenge was, it was a relief that we could see that, okay, it seems that we can do it.

Sheida Mirjahani:

But the second one is that this market is very rapidly changing. I mean, that you've seen that in 2023, there are a lot of either AI startups coming to the market, and also the AI infrastructure is growing like insanely fast. So for us as an AI company to keep up with the latest things, and also it lowers the barrier for new startups to come into the market. So now we have some competitors that they're not doing the same thing that we're doing, but the bar for us now is so much higher because we have to develop so much faster. And it's both interesting and challenging at the same time because it's like a game that you have to outperform, but still it's a challenge. And also, the other thing is that because there are so many options, we have to put a lot of time in researching to see what are the best tech tools, tech APIs that we can use to make a better product. So that one is also a challenge that we have faced.

Demetrios:

Yeah, that makes sense, especially because it's moving so fast. You're thinking, okay, we made the right decision, we're going with this tech stack, and then a week later you see some paper come out on Twitter and you go, wait, did we make the wrong decision?

Sheida Mirjahani:

Exactly.

Demetrios:

If you follow that, you can get that analysis paralysis and just never make a decision because you're thinking, well, if we just wait another week, maybe things will be better. And then too many weeks go by and you end up stalling.

Sheida Mirjahani:

Exactly. Yeah, that's what's happening almost every week.

Demetrios:

So one thing that I'm fascinated by is the different ways that cultures work. And I was part of this student organization when I was in college, and we would organize these internships in different countries so that the students could get experience working in different cultures and understanding how different cultures worked. I would love to know for you, you've been in Iran and now you're in Canada. What are some things that you've noticed about the cultural differences? I guess maybe you haven't, because you're the one that has your company in Canada. So maybe you would say, well, if anybody's feeling the cultural difference, you should ask my employees. But I can imagine there's some things that you've felt, and I would love to talk about that for a moment.

Sheida Mirjahani:

So if you're talking about the differences that I have experienced in the workplace, I would say that for now, it's been a few months that we have been working at a tech incubator called DMZ. It's a part of TMU University, which is in downtown. So this is the first time that actually, most of our team, the majority of our team are working together in person. And we are in a co working space with a lot of other founders. And I'm seeing that actually, there are a lot of similarities here between founders who have come from different backgrounds, but because we have experienced the same way of working, I mean, we have built something, we are marketing it. And, yeah, in the workplace, I haven't seen that much cultural differences, but of course, in my personal life, because there are so many cultures here, and I have found friends from all over the world. I have my latin friends, I have my friends who have immigrated from Europe, from Asia. So we have some funny moments together.

Sheida Mirjahani:

That, for example, I come from Iran, and whenever we meet each other, we handshake. And here I've seen that whenever you handshake with someone, they say, okay, we have just met. Like, why are you handshaking with me? And they either hug each other or do nothing. Every time I come to a workplace and I want to handshake with someone, I have to remind myself that, no, you have to hug that person or do nothing at all.

Demetrios:

That's classic. Yeah, it's like, do I handshake? Do I hug? Do I just do nothing? That can make for some fun moments, I imagine.

Sheida Mirjahani:

Yeah.

Demetrios:

Well, incredible. Is there anything else that you want to mention before we jump?

Sheida Mirjahani:

So I also wanted to give a shout out to Deepgram. Because we're using Deepgram for the speech to text API. It has saved so much time for the SLA that we have for our customers, I mean, the turnout time of their clips. So Deepgram converts an audio file of 1 hour long into a text in just two minutes. So, because the majority of the thing that our product does is that we are analyzing a text based on the audio or video, and we're turning it into a clip. Right. So Deepgram has helped us so much through their startup program that we are in. And also the other thing is that we are using them for audio intelligence features.

Sheida Mirjahani:

They're very cool. For example, as a part of the process of understanding a text and create the building blocks of a very good and engaging clip, we are using Deepgram to do topic detection. So what Deepgram does is that it tells us, okay, this part of the text is about these topics and these keywords and the other part is about these keywords. And we are using that to understand what are the main topics that are being spoken in this podcast, in this YouTube video, and then mix and match them into eclipse. So, yeah, I want to thank you and Deepgram team for making it possible for us.

Demetrios:

That's so cool. Well, yeah, I'm very excited by all the growth that Flowjin has seen and I wish you continued success on this journey. I'll be following along.

Sheida Mirjahani:

Thank you. Yeah, I'm excited to see where it goes.