October 31, 2022

Max Lukichev, Co-founder & CTO of Telmai Inc

Max Lukichev is a Software Engineer and Researcher turned entrepreneur with over 17 years of experience building large scale distributed systems, including streaming analytics systems. Max used to work in Fortune 500 companies, early and late stage startups, 3 of which became unicorns (Veeva, Reltio, SignalFx). In 2020 he co-founded Telmai and focused on building SaaS product in Data Observability space.

Julian: Everyone, thank you so much for joining the Behind Company Lines podcast. Today we have Max Lukichev, co-founder and CTO of Telm.ai .Telm.ai helps the data ops teams reduce the time spent on detecting and investigating data quality issues. Max, thank you so much for joining the show. I'm really excited to dive into your background being you were in software engineering and research and then jumped into entrepreneurship.

I'm, I'm sure that took a lot of. A courage and a lot of different experiences to lead up to this point. So, so excited to dive into that. But before we get into the good stuff, what were you doing before you started to me?  

Max: Well, hi Julian. We're excited to be here. Thanks for inviting before starting this company, it's actually my fifth startup. I'm not new to the startup world. It's the first time I've founder. But yeah, you know, early in my career, I started very early in, in the bigger companies, you know, Hewlett Park or Intel. But then kind of my life threw me into the startups of my first startup.

Like, oh, more than 11 years ago, I was also a kind of founding team member. One of the first engineers. It didn't work out. But then I joined another one which was still, you know, kind of average size startup, five years old, about 200 people. This was US Systems and which was basically a rocket ship.

It started growing so fast went ipo year after I joined. I with a blinker of night. It was 4,000 people, then 2000, 3000. And it just it's great success story, the company itself. Yeah, and I decided after that, you know, I always like to, to build things from scratch. You know, wanted to jump another early stage startup.

So I joined Romeo was also a very tiny company back then. It was 2015 50 people. Couple years old. Also was amazing, amazing journey. Last year. Well, I reached the unicorn status with evaluation over one B. After almost five years at, well, I decided, okay, it's time to do something new and exciting.

Another startup which was about the same size, but 200 people, 300 people, a less it was signal effects. And again, yes, got lucky. Couple months after I joined it got acquired by Splunk. So this is how I went back to the big company. Worked a little bit in the bigger company, and at the time I was already leading engineering teams.

I was engineering manager, director of engineering. So after a while, I just decide, okay. I got to build something new again, . And I had this idea of you know, I always had, I always been passionate about data, finding problems in the data, data mining, mm-hmm . And it was kind of very niche of things to do.

And luckily through this journey, through all of those companies, I met a lot of great people with whom I founded the company.  

Julian: Yeah. Yeah. What, it's so fascinating going from, you know, large company then from startup to successful startup to successful startup and, and going down the line, it sounds like one wasn't so successful, but the o the other two or three were just incredibly successful.

What, what are some of the common factors that you think made those companies successful? Was it the product, Was it the team? Was it how they, how they kind of, was it their go to market? What exactly do you think, or a few things that you think led to the success of, of the startups that, you know, went to unicorn status or were required.

Is there any similarity that you saw being on the engineering side that you think made them successful?  

Max: Well, if, if there was a simple answer, , everyone would have a successful company. It's not as simple, but one thing. What resonated with me and kind of the principle I tried to live by. Persistence will pay off.

You know, you, you have to, yeah. Stay true to, to what you're doing. Kind of keep doing. There's ups and downs. Like one day, everything good the other day, it's terrible. Just keep doing it. Don't give up.  

Julian: Okay. Yeah. How, how are engineering teams different? You know, and, and I guess more of a, a granular question, you know, being an engineer at Hewlett Packard versus running an engineering team, is there some takeaways that you took from that experience that, that worked well in the startup and environment and what about.

The startup environment is just uniquely different. Obviously it's a smaller scale, it's maybe a little bit more accountability, but at some point that there's an engine running. What, what's the like distinctly different between your experience working at a big company versus running an engineering team at a, at a small company? What's different and what's similar?  

Max: I'd say. In a startup the most critical skill is not skill, but you know, personality is, is ownership, right? It's, it's not the hard skill itself because in a startup, you do things you never done right? You, you have no clue how to do, but you have to figure out and you have to own it, right?

Because there's no one to. To keep, you know, pushing. You was like, Oh, do that, to do this. Right. It's, it's all about each individual in the startup owning the park they're building and trying to do best every day. Just, you know, just do a little more every day in that sense, mm-hmm. , it accumulates over time so that what makes startup be much more dynamic, be much more agile.

Because when you build, you don't know what you're building. You're still figuring out, You don't know what's the product market fit. You don't know your persona yet. You don't know so many things you're trying to figure out. And the quicker you can do it, the more questions you ask, the more you will try to figure out yourself.

Okay? You know, what if, right? What else we can try? The more successful, the more likely you will be successful.  

Julian: Yeah. What is something, what is the best way being that you, Okay,  

Max: so sorry. Go ahead.  

Julian: Oh no, I was, I was gonna ask you know, being that you've been on both sides of, as a founder and as a engineering lead and team lead, what is the best way that founders can communicate what they want their team to build and do?

So in a way that. You know, is productive and successful because I think a lot of, a lot of times it's, you know, road mapping and we give this grand idea about, you know, what we want it to look like, what features we want it to have. But the organization of completing those tasks is, is like the tricky part.

Being that you're on both sides. What, what are like the best practices for a founder, you know managing and, and inspiring and motivating their engineering team to build something that they're passionate about, but also, That can be done in, you know, in the time that it needs to be done. Do you, do you have any advice for founders?

Max: Yeah. Well, from this perspective it might be a little kind of one-sided advice because I'm coming from the technical background. So for me it's very easy to speak to engineers. I was an engineer myself. Yeah. And advice here is just be involved in a startup. It's really critical. Yeah. You have ideas But you have to be involved.

You cannot just describe your idea quickly to someone and, and be sure that it will be implemented in the right way, the way you intended it to be, and so on. So you just have to be out there with people building it together even though you may not be writing code yourself, but just participate in these conversations, give as much input as much.

context is possible because, you know, we all work with, you know. Yeah. All those engineers are smart people. It's that context, which makes that idea to actually materialize in something, you know, working. Right. You know, why, how the users are gonna use it. Yeah. Why they need to, It needs to be done this way.

And the more of this background you build. The more ownership you will end in the end because people will understand what they're building and the less of the, you know, never do micromanagement. Just instead build that ownership into, into people, so you don't have to do that.

Julian: Yeah. Yeah. I love, I love the piece about being involved. That's so simple. But I think in practice it's difficult because it's on the edge of, of, you know, being involved in micromanagement and being completely, you know away from the product. But I love the involvement of it in building together and giving somebody the.

The, the background and reasons behind because a lot of that, you know, I think breeds to creative solutions that you allow your employees and, and your engineering team to make by understanding the overall, you know, goal or mission. Thinking about your experience moving from, No, go ahead.  

Max: Yeah, just want to edit one more thing. And it's very important to build that culture that anyone can tell. Your idea sucks. My idea sucks because you know, it, it may be, you know, I want people to be not afraid to call it out. You know, there may be a better option, but again, my goal is to give that context so the best idea can win. .  

Julian: Yeah. Yeah, yeah. That, that's fantastic. When you think about your transition from engineer to founder, what are some of the things that you needed to learn or some, some the learning curve, anything on there that in particular you weren't aware of going into being a founder? That, that now in hindsight, you would, you would tell, you know, your past self, maybe focus on these key few things because if you learn them more quickly, it'll be a smoother transition.

Is there. Lessons or anything as a founder that you had to learn that you wish you knew going into it?  

Max: Yeah. Probably the toughest lesson is especially if you, you spend a lot of time kind of in bigger companies, more well funded companies building stuff, you kind of set your brain to work how to do it best, how to do it.

Right. You know, you know all of those best practices and you will try to do the, the best engineering possible and stuff. And the startup, you have to remind yourself about the right balance. You know how to do it right, but you have this amount of money and you have to do it, period, right? So you have to be creative into where to cut, where to do it faster, maybe not as ideal and yet, Be just the right level of, you know, hold it perfect, but at the same time, not perfect, right?

Just finding that balance and keeping reminding yourself like you don't have time, you have to move fast, you don't have money, you have to, you know, use whatever resources you have, whatever is available to you. So don't try to over-engineer, do what's necessary, but don't over. Because likely in half a year you will, your system will look way different from now.

You will have an opportunity to do it right later.

Julian: Yeah. Yeah. Thinking about your experience going through the acquisition from Signal at your time, from signal Effects just, just, you know, as a curiosity what was that transition like? Like reestablishing yourself on a new team? Were you brought over together and like a pod or were you having to integrate into that that larger organization and.

You know, what, what were some things you could bring over and what were some things that you had to adjust to with the new organization?  

Max: So I think single effects was a little bit of an outlier here because Splunk gave us a lot of freedom initially and a lot of time to adjust because it was such a, you know, Team integrating into Splunk's world.

So we had tremendous amount of support and time given to us to adjust to the way Splunk is doing things. So I really cannot call out any, you know, there were challenges, but it was, it went much smoother than I expected, to be honest. So, In probably the, the right, you know, if you're looking for advice or from my perspective as who I was, my role as it was back then, just try to figure out as quick as possible as the new organization works and try to adjust to that and just be transparent with, be supportive to, to your team because everyone is going through.

Uncertainty and stress. A lot of things are changing, you know, starting from little things the way, you know, expenses were done to the bigger things like how, what's the career path and stuff for, for people, how the roles will be adjusted and, and stuff like that. You know, how the teams will be split and then moved between different orgs.

All you can do is try to figure out and, and help others at this point. Yeah. And I can say that, you know, we did it just right. I was really impressed with the team that everyone just pulled together and was very supportive back then.  

Julian: Yeah. Now, Mo moving towards, you know, what you're doing at Tom, you said, you know, you've always been interested in big data and, and kind of reading into you know, kind of what qualities it has or maybe some insights that it shares what.

You know, if you could, if you could expand on that, what about data kind of gets you excited, and then what are some insights that companies gain from understanding their data just better or at a higher level or, or at a more granular detail?  

Max: Well, data is an exciting topic by itself because there are so many things to talk about.

And when we are talking about quality of the data, just imagine. On average companies would have, you know, dozens of different data sources. Sometimes it's hundreds. They're buying this data, they are producing this data internally. For example, every interaction of your customers are creating records in your databases, and you have a marketing system.

You have your customer support systems, your buying data from data providers, whatever. Never I have seen a case when this data is. There's so much garbage in the data. It's just unbelievable. And the thing, the, the most unbelievable thing that people, I think it's just nature of people. We all think that we don't have that problem until we actually see , someone shows us.

Right. And yeah, the, the exciting part is tremendous potential of having the right data at the right. What impact it can have on the business. You know, making the right decisions you know, where to, you know, where, where the problem can be with your business, where to put more focus. Or on the other hand, you know, we, we, we can be talking about the telemetry data versus sales data versus marketing data.

You know, great potential of fighting, building analytics on top of. But all analytics is based on the assumption the data is of high quality. And if you cannot Yeah. Ensure the data of high quality, like this is a wasted effort.  

Julian: Yeah. What, what's, what's an example of garbage data?  

Max: Okay. You can have systems which are, for example, collecting information about your customer.

and things can go wrong. At any point, let's say information about portion of the customers wasn't updated on time, Outdated data. That's a good example of, of bad data. It looks right, but it just a year old. It's no longer valid data where you have. Just plain wrong information. So you have wrong phone numbers associated with people, wrong addresses, wrong states.

All of this makes a significant impact for marketing. Yeah, data about your customers important properties of your customers, you know, wrong employee count, wrong size of the company, right? It's how you are targeting your campaign. It directly goes back to the ROI of these campaigns. Back in Viva, we, we used to work with the pharmaceuticals.

We were building software for pharma companies. Mm-hmm. there it's even worse, you know problem in your data during the clinical trial make us huge penalties and huge delays in the rollout of a new drug, which counts like millions of. Right. Simple mistakes, you know, the wrong dates, the wrong references to your patients.

You know, wrong attribution on this data. Anything can impact the audit process and trigger the very expensive losses.  

Julian: Yeah. Yeah. What how does he, I guess think about the quality of data and, and it, it sounds like it starts all at the collection point, but then obviously once it's, you know, goes through the analytic process, it, it kind of, it, it kinda relies on the data already previously, but how does Tom guarantee or.

You know, identify ways that you can have quality data. Is it helping on the intake process? Is it cleaning it up once it's been recorded? What, what does tell, where is Tom involved in the process of collecting quality data and having quality data to do analytics?  

Max: Yeah. Tell in this case is helping on as, as left as possible to the intent, right?

The clothes to. Injection of the data as possible. So, and it helps by monitoring the, the data by monitoring. What I, what I mean is if you look at the data observability as an area itself, it's a very broad area. There are a lot of software, like new software companies and sometimes an established software companies doing something in the data observability.

And a lot of times it's about. monitoring, you know, the pipelines, you know how your transformation of the data, you know, if there's any errors, if all the data's been delivered to your data warehouse for your analytics you know, some additional checks on the data wasn't delivered on time and stuff like that.

We are looking at it slightly different. So we actually looking deeper into the data and looking into the quality of the actual data being delivered. Not only the pipeline health, but also health of the data itself. And what we are doing is also employing ML ai to. establish how the data should look like based on the historical appearance of the data.

You know, the formats, the you know, the shape of the data, the distributions of the data. If they start shifting to something abnormal, we can alert the team responsible for that. So they can, in a timely way, very quickly see exactly what are the problems, investigate them, and really freeing down the storm around.

from the alert to a resolution you know, dramatically avoiding the problem when it's the business owner who finds the problem and the data and calls back saying, Hey, like, this chart doesn't look right. Why Right. In the form of escalation. So we are trying to remove that escalation from this past and be more proactive in finding those issues.

Julian: What, what, Describe the, the current traction of the company. You know, you went through yc, which is, you know, an amazing startup school to kind of help, you know, I, I think it's touched so many different startups and successful startups and has a really methodical way of doing so. But what has. From that experience have you kind of gained, and, and now in terms of traction that you're seeing, who are you partnering with and, and you know, what is Tam kind of the next step of Tama and the next milestone that you're currently working on and, and work and who are the companies you're currently working with?

Max: So, yeah, it's a great question. First of all, yes indeed. YC is a very, it's a great opportunity for startups. I won't say exactly learn how to do things, but give a perspective of what's important and our partners from Yeah. We, we have both YC and the institutional investors and all of our partners are just a great support.

I think this is the best thing that happens to the company, having, finding the the right partner. Who provide you advice who can help you with even marketing or even sometimes even closing the deals by, you know, coaching you. For me, coming as a, from the engineering side, it was very unnatural and I didn't really know how to do sales, right?

Never done this. Having someone who can coach you how to close the gear, what things to do, right, the sequence, how to even project those things to happen, it's, it's great. And. YC and our other investors are great support in that.  

Julian: Yeah. Yeah. And who are you currently working with that you're excited about?

Max: Customer-wise? Yeah. Yeah, we just recently closed Data stacks, which is a great company. It's my, one of my favorite databases. Cassandra, they're, yeah, very interesting use case. We are working with few more companies. It's all enterprise, right? So we, and I would say right now we are working with kind of bigger scale enter.

More established companies. Mm-hmm. , again, just because of the stage of the startup, like every startup goes through the stages. Initially. You do things which are not scalable. So you do the top down sales, you do you know, lot of stuff from your network and do more complex use case that maybe you would like to do just because you are due.

Design partnership with companies and so on. So it only certain companies have an appetite for that. Right. It's it's kind of the stage you are going through to, to get to the broader market, basically to find your product market fit. Yeah. So, yeah, we are working mostly with enterprise companies across different verticals.

So we are not limited to only to, let's say, technology or to pharma. We actually work with technology, pharma product companies in marketing. So yeah, it's, it's a good variety. There's no focus on one. Yeah.  

Julian: What, what's the process with working, you know, obviously data you need for, for, I'm assuming for your product to work very well.

Just a little large pool of it to, you know, run its processes, kind of iterate on different pieces of the technology and the code that you're running and, and you know, it, it seems like it's, it's very. It needs a lot of volume to, to work how it's supposed to. But what's the process that people don't know about in, in regards to working with enterprise companies?

How long is that process do you integrate within their team? Is it a proof of concept that you do with them before you, you know, go through a contract? Is it a white glove service? What, what is it like, what's the experience of, of working with enterprise clients that is just ultimately different than, than working with, you know, small to medium businesses, mid-market or, or even, you know,

Max: Yeah, this is a great question. So in probably the short answer is whatever your worst expectation work with enterprise in terms of like how complex it is multiplied by three. Because it is brought and a challenge on every step. First, you have to find the right people who have the budget, who have the use case.

So it's the, the discovery and validating the, the poc if the POC is even worse. Starting. So after you do that, Yeah. The POC itself, there's a lot of upfront work you have to perform such. Do security review, do all the sign offs on the POC itself, even though it's unpaid poc, right? You have to make sure that you are engaging it.

Yeah. Knowing that there is all the things done, right? You get, you know, five different approvals from different teams to, to even start the poc. Once the POC done and poc usually with enterprise companies, they're also much more complex than bigger enterprise. I mean more sophisticated enterprise than the smaller size businesses is.

They're much more custom. Like a lot of times there is a unique requirement, a new integration, let's say database you haven't had in your portfolio before, or, you know, a new type of analytics that. There in the product before, you have to do it very quickly during the poc, develop something and roll it out to demonstrate the power of the platform or you know, the architecture that you have.

And after that it comes to the negotiation and that makes it even much more fun because the enterprises, they work on their own financial model. You know, The financial year is different. There's a lot of times, depending on when you entered in this financial year, was it initiative planned or not?

So you have to do a lot of financial engineering how to make sure that okay, even if you are good, if you want the technic, you have the technical win through the poc, how to close it for you know, to make sense for yourself and for the company, which is buying it. Because again, there are so many restrictions.

You know, the budgeting, the planning when this happens throughout the year cycle and so on.  

Julian: Yeah, yeah, yeah. It, it, it just seems like there's so much involved. Oh, good.  

Max: Yeah. Not to say that security is a must, like you, you have to have soc tool. for sure. Yeah.  

Julian: Yeah, yeah, yeah. It, it, it's, it's, I love the honesty behind, you know, whatever your worst expectation is, multiply it by three.

Because I think, you know, a lot of companies strive to work with enterprise clients because, you know, I think in their, our mind, the, the deal size goes up. But it, it just so much more involved of a process that a lot of companies either need to, to mature to, or. Or their product like yours needs to be inherently useful for those companies.

And, and it kind of justifies the, the level of involvement. What are some of the biggest risks that Telmai faces today?  

Max: I would say the, the biggest thing I worry about is just general market because for a startup startup, being in startup, it's always being vulnerable, right? And a lot of things are out of your.

If the market goes down, everyone starts shifting kind of cutting the spend everywhere, right? So it kind of creates this, you know, cycle. Like, okay, you, you have to have certain cadence of of the new customers. The revenue grows yet lot of bigger companies, especially, you know, the enterprises, they're trying to do their financial planning and trying to cut their cost to prepare for the longer pull down in the market and so on, right?

So yeah, depending on like in what shape you were entering this place of the market, it may be good for you or bad for you. It all depends like what's your monthly. , you know what's your projections in in the pipeline and the growths? How big is your team? Right. Luckily for Temi, we were, we didn't have a lot of kind of this ballast.

So, and, and a lot of kind of in a lot of spend. So we are a little bit feeling a little bit more comfortable in this sense. But, you know, if you. This stage or where we all are now, and you have, you know, 200 engineers and not that much of revenue, that's a big problem. You, you have to be super worried and

Julian: Yeah, yeah. If everything goes well, what's the long term vision for Tom? Tell,  

Max: Long term is really growing it into a complex solution for. auction for all things related data quality, right. And monitoring data quality. I don't think we will go into the infrastructure like data engineering infrastructure is not exactly where we see our ourselves, but, things about the data quality what's inside the data.

That's definitely our.  

Julian: Yeah. Yeah. I, I always like to ask founders this question just to gain some selfish research purpose for some selfish research purposes, but also for my audience to gain some knowledge and insights. But what are some books or people who have influenced you the most throughout, you know, your career or currently as you're building a startup?

Max: I think. You know, I was fortunate enough to work with some great leaders through my career and I even picked up some of the, you know, mottos which kind of helped me moving Yeah. Day to day. And one of that was Peter Gasner from CEO of Viva. That. Persistence will, will pay off is one thing which kept me going in a, in a very hard moment.

Lot of times that's his phrase, which was from the water cooler when I was a young engineer back then. .  

Julian: Yeah. Yeah. I love that. I love that. It's so simple and, and it and, but it, it, it just is empowering for the amount of work that you have to do. And, you know, running a startup now myself, it's, it's really, you know, the achievements come.

the re, you can't rely on that as the measure of successes. It's the work that you put in, that's your measurements, the activity that, that you should, you should help validate the work that you're doing because at some point you should see a lot of success and things trickle in, but it's all the work that you're doing and, and prior to that, that really leads to those moments.

I know we're at time. So thankful that you jumped on the show. But before we leave, where can we find, where can we support? Tell ai give us your LinkedIns, your Twitters, your every handle. If you have a Discord channel, where can we be a supporter and and, and, you know kind of dive into technology that you're building.

Max: Well, first of all, thanks so much for inviting. It was a great chat. I really enjoyed the conversation. You can learn about Telmai on our website telm.ai and also on the, our LinkedIn page just by looking for tell my T E L M A I, just as a single word a little history behind that. Tell my, actually just means tell me ai.

So that's what we are doing. Tell me where is my problem. AI please.

Julian: I love that. Well, thank you so much Max, for being on the show. I really hope you enjoyed yourself and I'm, I'm really excited to share this with our audience. And maybe in the future I hope to have you.  

Max: Thank you so much. Have a good day.  

Julian: Yeah.

Other interesting podcasts