Jake Aaron Villarreal: I'm Jake Aaron Villarreal, born and raised in Silicon Valley, and here to take you behind the scenes to share what it's like to be a startup founder, the journey they're on, the problems they face, the products they build, and effort to make our lives better. I'm excited to have with us today, Dylan Fox, founder and CEO of AssemblyAI. Dylan, welcome to the show.
Dylan Fox: Thanks for having me on here.
Jake Aaron Villarreal: Well, Dylan, I'm excited to have you on. I know we spoke a few months back and have been seeing how your company's progressed and in the news. And really a lot of the companies that use your product underneath as the infrastructure, the APIs behind their product that we see and hear about all the time. Um, before we dive into your background, where are you joining us from today?
Dylan Fox: So, I know you're in uh Bay Area. I used to live out there about eight years and then um now live in Brooklyn, New York with my wife, our, our two kids. So New York City based.
Jake Aaron Villarreal: Hundreds of AI startups are launching every month battling to build their founding teams. As a leader, your job is to get results. When it comes to hiring, that's where it gets tough. So you go out and you try a recruitment firm, but they don't understand your story. They're off target and when they send you candidates, it's a waste of time. We believe you should never have your time wasted. That's why we launched Match Relevant, because your story is more than just an open role. It's your founders' journey, the problem you're solving, the product you're building, and why it matters. When we work with companies, we make sure we understand your whole story. So, we go out and do a search, we're on target, it's worth their time, they're interested, and more importantly, it's worth yours. And when it comes to hiring engineers, we work to make sure we get it right by deploying a team of seasoned CTOs that have built some of Silicon Valley's best companies. They can collaborate with you in the technical interviewing process. They can be a sounding board or they can run it for you. When it comes to building teams, there's no time to waste. Let's make it count. If you have a role that needs to be filled, book a time with a hiring guide at matchrelevant.com and learn how we do it.
Very cool. One of my favorite places to be and I spent some time out there myself. So, just it's hard to uh beat that energy and the vibe and tech and everything's there. So, well, thanks for joining us.
Dylan Fox: The tech scene definitely has like exploded here over the last couple years. I definitely have, have felt that compared to when I visited here, you know, like a decade ago. Um, so yeah, it's really, really uh really great city.
Jake Aaron Villarreal: Yeah. Well, cool. You're in the right spot. Um, a little bit more about AssemblyAI. It's a leading company that has raised 115 million from Accel, Insight, Smith Point Capital, and Y Combinator to build AI systems that transform human speech into meaningful outcomes and product experiences. Dylan has a background as a research engineer, lives in Brooklyn, New York, as we talked about, and excited to talk a little bit more about what they're doing in this world with AI coming into all sorts of every industry and all businesses. Um I guess before we do that, Dylan, give us a little background of you. Um your origin story, how did you really start off and get into research or technology? And more importantly, yeah, what, what, what was the, the inspiration to become an entrepreneur?
Dylan Fox: Yeah. Yeah, for sure. So, I um in college, you know, and I don't really remember what prompted me to do this, but um entered into this uh uh new business plan competition that the, the, the university I went to put on each year. And you, you know, wrote up a business plan, which you actually had to write because this is like prior to, you know, ChatGPT, of course. But like wrote up a business plan. Um you submitted it, you, you, if you got selected like presented, and then uh the goal was like if you placed in the top I think it was like three or four, like you would get some, some seed funding of like a couple thousand dollars to like get started. Um but I entered into that uh new business plan competition. Thought it was really fun. Uh I kind of loved the idea of just like thinking about starting a company. And after that uh did start a couple of like random startup companies with some um uh friends in, in college that I, I went to college with.
And through that experience got into programming and software development. So I, for a couple projects that, or startups we were working on, like had... you know we had to like build an app. So I bought a bunch of programming books, started to learn how to code, um and fell in love with software development and programming and building things. I think like for me I really like building things whether it's like product or software or company, I find that um uh challenge of, of, of, of building things like the, the, the really fun. Um, as I got into software development, you know, I kind of kept pulling on that thread and that led me to like kind of like went through the stack, and started with like front-end development and then back-end development and then machine learning. And then really got into machine learning in like 2012, you know, to '15 where it was still mostly classical machine learning based um techniques that were being used in the industry, but really wanted to like focus on hard technology problems is like where I kind of gravitated towards.
[Dylan's Entrepreneurial Journey]
So, um, ended up taking a job out in San Francisco at Cisco to work on this, um, research engineering team. And that's where I just like really started getting into, you know, deep learning, AI, neural networks, and here we are. Here we are now almost, almost 10 years later.
Jake Aaron Villarreal: Yeah. God, there's so many companies that spin out of Cisco and Intel and some of these big uh, hardware and software companies. Um, and AI, it's been around for so long, but it doesn't... hadn't really hit the market or the consumer market as we know it today. So, it's really sounds like you were positioned well before everything kind of came to the surface. Um...
Dylan Fox: Yeah. Yeah. I remember like neural networks, you know, were starting to get popular in that 2015, '16 time frame, but most people were still using classical machine learning systems in production and folks were starting to switch those to neural network and deep learning based models. But it was just so clear that like these neural network uh based models were going to be so much better and the, the sky was like, the, the ceiling was like nowhere to be found. So there's a lot of runway and it was just going to completely change a lot of different you know industries and the way that machine learning systems were built. So um yeah I remember going to like the first TensorFlow meetup that they put on. This is when people were still building their models in like, in, in um in TensorFlow.
Jake Aaron Villarreal: Wow.
Dylan Fox: And uh like was you know training models on these Nvidia K80 GPUs which now you know are like dinosaurs compared to what's available today. But yeah, the industry is just like, the pace of, of progress is just insane.
Jake Aaron Villarreal: Yeah, every, every day it seems like you learn about some new model or some new acceleration that's helping companies do more with less. And infrastructure-wise, you know, costs seem to be somewhat going down. It's still very expensive, but um it's cool to see that happening, that innovation. Let's talk a little bit about AssemblyAI. So, what was the idea that prompted you to want to start Assembly and really what, what's the problem it solves today?
Dylan Fox: Yeah. So, we're focused on building the industry's best speech AI technology and then making it super easy for product teams and developers to integrate into their products. So what I saw back in, this is like 2017-ish time frame now, was that this speech AI technology for tasks like speech-to-text, speech recognition, speaker identification, you know, any, anything related to speech processing. Um, the technology was just getting so much better, but unless you had like a huge team you could, you could uh build out to work on it internally, like you didn't really have access to the latest and greatest technology. And so the idea we had was, "Hey, let's really push the frontier of this technology and, and what it can do for people that are actually trying to build applications with it. Like not just for academic purposes but like, you know, to get it out there in the real world and solve those real world problems. And then let's make it available through a super simple developer API like Twilio or Stripe, so that anyone, whether you're a college student or big um software developer at a big enterprise company, you just sign up, start building, put into production without having to talk to anyone."
And make that technology like really easy for people to build with, because the types of applications and products people could build when you, you know democratize this technology is, is really exciting and is really cool. And we see that now. You know, we have like tens of thousands of people using the API for all different types of products and applications and they're having you know amazing impact for end users in the real world. Whether it's like helping people learn, or helping salespeople um uh improve their craft, or helping in recruiting um you know people like uh um automate parts of the recruiting process to, to be able to meet more candidates. Like there's a ton of applications for this tech and there's still so much, so much room to go on making it better. So that's, that's how we got our start. And you know now where we are today is we've got yeah tens of thousands of people using the API, um some big companies that we work with in like contact center space and um media and in virtual meeting space. And we uh a lot of great new products that we're launching as we like continue to continue to grow.
Jake Aaron Villarreal: Yeah, that's great. Can you share some of the company brands that we would know and some of the use cases? You've mentioned, you know, recruiting in some different areas, but what are, what are some of the big major ones that maybe we wouldn't know that your technology is really supporting and it's really been a success?
Dylan Fox: Yeah. So, you know, what I love about our company is that there's been so many new AI-native startups and new companies that have come to market over the last two years as LLMs have um gotten better, as just AI technology across the industry has gotten better. And so every day we're seeing really innovative new startup companies come to market and then just like take off and grow super... in the software development space, companies like Cursor and Lovable that like just are blow, blow up and grow so fast because there's so much demand for what...
[The Birth of AssemblyAI]
...they're building and they're taking it to market really fast. And so for us, what I think of are companies like Fireflies, which like, maybe you've used them, but they're AI notetaker, got millions and millions of users. They're growing super fast and they have an AI that can take notes for your meeting. They are um uh in more meetings from like an audio duration, some... than Spotify is like generating audio, which is crazy. Um, and the, the adoption they're seeing is just like mind-blowing because they can, you know, their AI can like help you take notes in your meetings, create action items for you, uh, create like great summaries for your team. Um, and they're one of, you know, thousands of startup companies that are building really innovative products with our API and with our, our, our models. Um but then we also work with companies like Zoom um and you know more traditional enterprise companies that are, that are earlier in their journey of deploying AI. But you know we have bunch of really exciting um uh like customer stories on that front we hope to share in the, the coming weeks and months.
Jake Aaron Villarreal: Really cool. Well, Fireflies was actually the CEO was on our podcast a while back.
Dylan Fox: Oh, nice.
Jake Aaron Villarreal: Yeah. And uh you know we, I, I think most companies today use some sort of a, an application that can record meetings and summarize it and give intel and ask questions to it. But it's so incredibly accurate. It's, it's mind-blowing.
Dylan Fox: And yeah, that's one of the fastest, you know, growing use cases that we support is what we think of as, you know, conversation intelligence. So it's just an- analyzing and being able to capture conversations for a variety of use cases. You know, another example I think of is a startup company called Xero, and they have an AI that if you're in field sales, you can, you can take with you as you're going from house to house and it can listen in on your conversations out in the field. It can give you tips after the conversation to help you improve as a salesperson. And they're helping sales reps earn 10 to 20K more per quarter in take-home pay because traditionally coaching a field sales rep is really hard. You have to like go with them, listen in in person. It's super awkward for everyone. But they have an AI now that can do that, and it's really helping people um improve their, their craft as a field salesperson and Xero is growing super fast as a result. Um they just had this big partnership with ServiceTitan. So there's so many examples like this I could give you where these really innovative new...
Jake Aaron Villarreal: Yeah. Incredible. You know, as you start a company, it's really about first having a product that's got a product-market fit and then going out and having customers use it, or potential prospects adopt what you have to sell. What was your first big win? If you can recall, when your product was received, used, and you thought, "That's a great logo" or "That's a great company that can really help build on how we're getting our product in the market and awareness for what we do."
Dylan Fox: Yeah. Well, I think, you know, one of the things I've learned is like product-market fit is like a journey, not a, a destination. Sounds super corny to say, but I think you're, especially when you're building in a very dynamic market, like we are in AI, um the market is changing. So, your product-market fit always is changing as a result too. And you're always trying to optimize that, that like product fit with a dynamic market. So, um, that's was something that we love as a company is how do we really focus our R&D and what we're building and our innovation like on customers and not just on the technology or internally, but like on our customers.
Um but to answer your question, you know, a few years ago... so we started the company in 2017, and then we spent the first three years just like building the product. Um super small team, and then launched it around like 2020. Um, and one of the first customers that I remember that deployed us into production was this company um, called CallRail, who's still a customer now many years later, and we love them. And they have, you know, over 250,000 small businesses that use their product to um help record and analyze call conversations into their small business and help those small business owners better understand where calls are coming from, how people are hearing about their small business. They can, you know, have some like um intelligence on what people are asking about. So, it's really cool application.
And I remember, you know, I was still, you know, manning our like support chat at this time, and their CEO like wrote into the support chat. And, you know, we got on the phone with them and they were really excited about what we were doing. And they used to get a ton of complaints on quality for anything like speech AI related, speech-to-text related in their product. And then they switched and like all those complaints went away and their conversion rates like you know have like 2x, and you know they, they, they're working on a ton of really innovative stuff now. Um, but that was definitely a cool milestone for the company when we got to a point where there were companies like that that were starting to put our, our, our product into production um at pretty large scale. And from there, you know, it was just, you know, more and more of those. And um uh yeah, here, here we are. Here we are now.
Jake Aaron Villarreal: Yeah, that's great. You know, there's a lot of different ways to go about getting your product into the market when you're presenting or selling or pitching your product. Who cares most within a startup or a company that wants to hear your product? Are you selling to engineers or is it the founders or is it a combination? What's that look like?
Dylan Fox: Yeah. So I think for our company um really it is like a mix of product...
[Innovative Use Cases and Customer Success]
...leaders and engineering leaders, and now at smaller companies those are the founders sometimes, at bigger companies those are also the founders. I think the cool thing about what we're creating is that we really are like one, critical infrastructure, and then two, a pretty strategic like technology decision that these companies, you know, are making. Because we meet with customers all the time that are like looking for our advice on how they should be architecting their infrastructure so that they can move really quick, innovate, um, ship new features really fast, because everyone is optimizing for speed right now. Speed to market because the market is so dynamic. There's so much demand out there for all these AI features and applications that everyone's trying to like get to market really, really fast and keep iterating really, really fast. So um you know we, we are often working with the product leaders, engineering leaders um and partnering with them to help them figure out you know how to build this speech-driven AI application or workflow they're trying to build and then how we can best plug in and support them to do that. And like what features and capabilities of ours are, are the best for them to leverage depending on what they're trying to do.
Jake Aaron Villarreal: Yeah, that's great. You know this transformation within AI we're seeing a lot of building of uh AI agents, agentic AI. And um you know our own company, we recruit, but now we've deployed agents that are helping us do our job better, faster um and be more strategic in how we use our time. Uh you know I, I kind of see this world that others have already talked about, which are you know agents talking to other agents to get work done. And it may replace people, it may make others better at what they do, might change you know the dynamics of our workforce. But um what are the trends uh in how people are building agents and workloads from your perspective?
Dylan Fox: Yeah. So what we saw early was that you know because we have a really interesting point of view with thousands of like application developers building with our products, with like front row seats to like what people are trying to do and where they're going. Um and what we saw, which is now a pretty mainstream design pattern within AI workflows, is that um application developers are really leveraging a mix of AI technologies and best-of-breed like AI models and technologies in a coordinated fashion to create these really reliable, accurate, like cost-efficient you know whatever dimension you're looking at, AI workflows, agents, whatever you want to call them.
And so if you really were to look under the hood of like any of these agents or AI-native products, the infrastructure looks like... if you were to, an analogy would be like if you look at any modern SaaS application and or any you know technology application, you've got a bunch of different um technologies in there, right? You've got AWS, you've maybe got Datadog, you've got Stripe, you've got Twilio, you've got um you know pick, pick your, pick your like part of the stack. And you've got some like AI application there, or sorry, some AI infrastructure there. You're seeing that within AI too. So you know you've got reasoning models, you've got maybe smaller models for different tasks, you've got specialized models for tasks like speech related or image related or video related, and then you're coordinating all these together in this agentic workflow or this design pattern that's reliable and that's interpretable, and that you can go in and tweak and make edits to to put guardrails in and actually get this stuff out into production.
You know, when you look at enterprise adoption of AI, um there's still a lot of like... we're still actually pretty early in that. Uh if you look at any of the reports from McKinsey or Bain or whatever, um because people are really thinking about inaccuracy, reliability, um data security, and so they don't want to just like put black boxes out in, into their companies or into, you know, with, with, into the hands of their customers. They really want to build these systems that like they can go in and have control over. And what's amazing about these agentic workflows and these agents is like they provide that level of interpretability and ability to go in and like you know make changes and edits that um people deploying this technology like want to have. And that design pattern has been what's been a huge enablement of actually being able to get this stuff out of into production.
Whereas like initially people were just taking one model, trying to get as much as they could done with that. And you could get a cool demo done, but then you'd put it in production and like all these issues would come out. And now I, I think of the design pattern we're seeing in production now is almost like the 2.0 you know of that, which is what we're now calling agents or these workflows where you're really able to pull a bunch of different AI technologies together that are now all more mature... a bunch of different models, but you know observability, vector databases, whatever... and build these reliable like robust um workflows or agents that can go out there and like you said, for recruiting or sales or whatever task, um research, actually drive like a lot of success.
[Product Market Fit and Early Wins]
Um, and we, we use a ton of AI tools internally, too. I mean, we have AI workflows that are constantly summarizing customer feedback and trends and putting those into the hands of everyone at the company. There's so much, you know, power that you can leverage um these AI systems for and it's, it's really cool to see.
Jake Aaron Villarreal: Yeah, it's incredible. Yeah, just the pace. You know, I was reading the Wall Street Journal a few weeks back and I read an article about uh the automotive industry and they were talking about, you know, it's not the big that's going to kill the small, it's the the fast they're going to kill the slow when it comes to technology and innovation and in AI and all different aspects of industries. Um, you know, staying ahead of the competition uh in this market's tough. Um, if you were to look at a company like Google or OpenAI and if they open sourced better speech models tomorrow, or...
Dylan Fox: Yeah.
Jake Aaron Villarreal: ...you know, did something that was transformative, what would, how would that impact AssemblyAI um from your perspective, if at all?
Dylan Fox: Yeah, it's a, it's a great question. You know, something that we talk about a lot at the company is like we want to be the most customer-focused AI company like on the planet, right? So, a lot of AI companies are focusing on improving...
[The Rise of AI Agents and Workflows]
...technology to improve the technology and to improve certain benchmarks or whatever that's actually like pretty untethered to what customers want or care about. And that's why I actually don't really think like foundation models as you would think of them are products in their own right, because they're not really built for anyone or anything. They're just like cool technology primitives that you can leverage like any other technology primitive to go build a product or a service that you can put into market. But on its own, it's just like a piece of technology and then the work is like packaging that up into a product or service that you can take to market. And so I think because the market's so early, um, people are still figuring out, you know, how to like wade through all this.
But for us as a company, what we think about is we're focused on creating models that are perfectly aligned with what the customers in the markets that we serve really care about and want. And so, you know, for example, we're launching new models for voice agents. And all of the eval, eval metrics and eval data sets that we really try to optimize that model for are like perfectly aligned with all of the feedback that we get from people building voice agents on the front lines and the failure modes. They actually, you know, the models that we're putting out might be worse on certain academic benchmarks than other models, right? But we don't care because that's not what we're solving for. We're not solving for leaderboards. We're solving for like what do our customers want?
And then in addition to the models, what we also really focus on are what are all the things around them that you need as a product team, as an application developer to quickly get these things into production, um be able to iterate really, really quick as you're getting customer feedback. So what's all that, what's all the, the functionality and capabilities around the, the core models that you really need? I think of an example, a stat is like something around like 70% of all the traffic that we get to our API has features like language detection or speaker identification turned on. And those are like critical capabilities that application developers rely on, and we go really deep there for our customers just like we do on the core speech-to-text and speech technology.
But then the third thing we really focus on is you know uh making it really easy to deploy this technology anywhere so that you can be really, really fast with getting stuff to market and meeting all the data security needs you have, or meeting all the um security and compliance needs that you have. So you said it right. It's like the fast versus the slow. It's not the big versus the small. And what we see time and time again is like our customers are able to move so much faster because they're able to leverage our infrastructure and all the work we're doing, almost operating as like an outsourced R&D team for our customers. And they can focus on building the products that they need, iterating super fast, what, where they're trying to push their road maps. They feed that back to us, and then we go and we build that into our road map for them and give them better and better technology all the time that helps them push into new markets, make their products better.
So, I would say overall we're, we're very customer focused with what we're building, which is a shift from how we even thought about the techn-, the product that we were building a couple years ago where we really were... the technology was in a more uh early stage. Really were focused on just the technology and making it better, but now we're really focused on what are people using this stuff for, what do they really care about in the markets that they're using this stuff for, and how do we just, you know, solve all of their pain um as much as possible. And that's how we differentiate from a technology and product perspective.
Jake Aaron Villarreal: That's a question that always comes up when we're talking to companies, but uh I like... I'm very pro-AI. I see the benefits of it and I, I love the direction where the industry is headed and the innovation that's coming out of it. Um, so I think you're in a great spot. When you, when you look at your company today, what's the biggest challenge you see as you continue to go through 2025 and beyond?
Dylan Fox: Yeah. So, we...
[Navigating Competition in AI Technology]
...like everything, take a very customer-focused view on. I mean, obviously there's like a million things to always um solve and fix when you're building a startup company, but like I think for us, it's all really in service of like the customers that we're trying to go um support. And so, you know, I, I think there's like two things that we really see as company that we're focused on. So one is uh these voice agents that you're seeing in the market are starting to be really successful from a success rate perspective. You know, B2B voice agent conversations are having like upwards of 40, 50% success rate. And wow, that's, that's pretty, you know, that's pretty good. Um, and you're seeing rapid, yeah, rapid deployment of, of these AIs to handle phone calls.
I, I go back to that customer example, CallRail. You know, of the 250,000 SMBs that they, that they have as customers, um you know those SMBs get a lot of phone calls after hours that no one answers, or, or they're, you know, you're calling into a shop and it's one person on staff and that person's helping a customer, and you know, no one answers the phone. So, even just being able to have an AI that can answer the phone, um, answer people's questions, give them, you know, information on store hours, uh, take down their information for a callback. That's super powerful and there's a lot of value in that. And so we're launching a bunch of new products for um uh developers and companies that are building voice agents that will make those voice agents more successful, faster, more accurate, especially in like key pieces of um uh key areas where voice agents fail today. Like being able to tell when someone's done talking, you should have the AI respond or interrupt or whatever. So we're actually a few weeks away... I mean it's end of April now. I don't know when this is going to be released, but like we're a few weeks away from launching that product which we're really excited about, and we have like a huge waitlist for it. So there's just, you know, um ton of interest in that.
And then you know I would say the other is what we talked about earlier. A lot of com-, a lot of products and workflows built around conversations, whether it's a podcast or a video or a sales call or a customer support call or internal meeting, whatever. Um, today they lack a lot of context and so there's some limits in the, the, the accuracy and the, the out-, these new models and products that can take in a lot more context and be a lot more accurate at runtime with really key pieces of information like product names or people's names or you know medical terms. You, you could think about it like if a, give a medical AI scribe, and it gets the prescription name wrong or the dosage amount wrong, it's pretty bad, right? Um and right, like models today struggle in those areas without prior context knowing like, "hey, this is a heart doctor and the patient came in for these reasons." That information and context really helps these models. So we're launching actually this week our, our next generation model that can take in that context, be much more accurate at inference time in really key areas that help these agents and AI workflows around speech be much more successful. So, we're really excited about that, too. Um, to solve these problems around like context, accuracy, uh, to help our, our customers and people building these speech AI-driven applications to, to drive a lot more successful outcomes.
Jake Aaron Villarreal: Yeah. Well, we've got a lot of clients that are currently building AI voice agents as we speak, and we're helping them hire engineers. And, you know, talent for that skill set's, you know, hard to find. I will tell you firsthand.
Dylan Fox: Yeah. Yeah. Um, totally.
Jake Aaron Villarreal: But, uh, the demand is out there for, uh, you know, a lot of companies building in that space. So, I think you're, you're on target with a lot of companies that are carving out their own little niches. Um, what's worked for you as a company? You've got 100 employees and growing. Like, where, where, what's worked for you in terms of finding the right people to build your company and to continue to scale at the level you guys are?
Dylan Fox: Yeah. Um, I think the biggest thing, I'll take that back. I think one of the biggest things, because there's a lot of big things I've learned. One of the biggest things is like, and this is for other founders that are maybe earlier in their journey, is like um startup companies, building a company is not like building a franchise business. And there's a lot of conventional wisdom out there that is helpful in like how to build uh a team or how to build a company or what departments you need or whatever, just anything related to building companies. Take that in. But I think you know, designing a company that...
[Customer-Centric Product Development]
...is like, has product-market fit with your customers is like really important. So like, how do your customers want to buy? How do they want to get supported? Um how do you want to uh deliver like product and feature updates to them? Like what are the important things for you to really focus on? And doing that, like bu-, building your team and building your company in a way that is like really aligned with your customers I think is really important to do. And you know, probably some of our biggest pain points have been when we've tried to build a company you know following like a traditional playbook that just like wasn't a good fit for our customers at the end of the day. And so there's friction, and you know, you want to like be... you want to remove like all possible friction in your, in your, in your uh, in your company and like how you're building, right? So um, that, that's been one of the biggest things I think that we've learned as a company and that I've learned as an entrepreneur is like, yeah, really building the right team and structure and company and process, whatever, for your customers is like really powerful.
Jake Aaron Villarreal: Yeah, that's a great insight. You know, a lot of companies uh have been remote, and now they're moving to hybrid and now they're moving to on-site. And uh a lot prefer to be remote and distributed. What's your position on that?
Dylan Fox: Yeah, so we are a remote, distributed company. We do get people together a lot in person though, because I do think it is hard to build trust you know as like colleagues like fully remote. Um so we do like, we do really value in-person time and we do get people together a lot in person. Um we do like full company off-sites twice a year. We had one in January. Our next one's coming up in a couple months. And those are awesome, they really help people like build relationships. And uh that being said though, like we are um remote company like, plan to stay remote. I think there's, the pros outweigh the cons of being remote, but you have to kind of like actively think about the cons and how you try to mitigate them, because they don't just kind of like disappear. And you've got to, you've got to really manage them well. Um, and I think there's like skills that people and companies can develop as remote companies that like are competencies. So like how you communicate in Slack, openness, transparency around stuff. Um, that just help make it easier for, for people. But yeah, we, we, we think the, the, the pros outweigh the cons, but in-person time is definitely like still super valuable and we, we do prioritize it, especially around like key meetings or key like planning cycles or whatever, where we feel like, you know, this would just be easier to have like this five hour session like in a room together um not like on a Zoom.
Jake Aaron Villarreal: Yeah, I hear you. Yeah. Well, I know you're growing. Are there specific kind of roles? If you're an engineer out there listening, maybe you're in the US, you're in Canada, you're South America, wherever. What uh where are you looking to grow? So, if people are interested, they can apply directly to you. They can um know what you're looking for.
Dylan Fox: Yeah. Um you know, we are handling pretty large scale now with our API. So I think it's something like you know tens of terabytes a day of voice data that we're handling through our API and that's growing really quickly. Um it's up like you know almost 3x year-over-year. So, um, we're always looking for more engineers that can help us build really resilient, scalable, um, inference infrastructure to meet the...
[Addressing Security and Privacy Concerns]
...demand of all of our users and customers that we have on the API. Um and then engineers that are really interested in helping to figure out how we continue to package up all this technology in more and more um uh kind of packages that can be run anywhere, deployed anywhere. Um so those are two areas where we're hiring for on our engineering team right now. Um, and then, uh, on the marketing front, um, we have some exciting open roles too on our, on our marketing team, uh, for folks to check out if they're interested in that.
Jake Aaron Villarreal: Very cool. I love that. If anybody wants to find you or find AssemblyAI or maybe find the roles, where would they go?
Dylan Fox: So I, if you go to our website assemblyai.com, uh there's a careers link all the way in...
[Challenges and Innovations in Voice AI]
...the footer. So check that out. Um and then for me, uh LinkedIn is probably the best place to you know, if you want to reach out or say hi or whatever, my name on LinkedIn, look me up and um that's where you'll find me. I don't really... I, I browse Twitter/X, but I'm not really active on there. But, but uh LinkedIn's, you know, would, would be the place.
Jake Aaron Villarreal: There you go. Perfect. Or if you're just walking down Brooklyn, some side street, maybe you'll see you in a cafe or something. You never know.
Dylan Fox: Yeah. Exactly. Exactly.
Jake Aaron Villarreal: Well, I want to thank you so much, Dylan, for coming on and sharing your journey, your origin story, and really what you're bringing to market and have brought to market. I think it's phenomenal. I'm sure we're using your product at some level. Um, and I have to tell you, it's great. So, um, thanks for coming on and thanks for the listeners for listening. It means a lot to me that you spent your time with us today. I'm your host, Jake Aaron Villarreal signing off for now. Can't wait to catch up with you all in the next episode. Until then, Dylan, the world, take care. If you like what we're doing, don't forget to subscribe, leave a review on Apple Podcast or wherever you listen, and follow us on YouTube where we go behind the scenes to learn what it takes to be a startup founder.
.jpg)
.webp)
.webp)
.webp)
