Laura Ellis @ Rapid7 | Data Science Hangout
We were recently joined by Laura Ellis, VP of Data at Rapid7 to chat about making artificial intelligence and analytics accessible to everyone in a secure and scalable manner. Speaker bio: Laura Ellis is the Vice President of Data at Rapid7, a cybersecurity company. Her mission is to make artificial intelligence, data science and analytics accessible to everyone in a secure and scalable manner. She has worked in the data field holding a variety of positions for almost 20 years. She holds a Bachelor of Software Engineering from the University of Western Ontario, a Master of Science in Predictive Analytics from Northwestern University and a Chief Data Officer executive certificate from Carnegie Mellon. She is the co-founder of Data Mishaps Night (www.datamishapsnight.com), an evening for data practitioners to share their data mistakes and learning. She has a long time blog where she writes about all things data: www.littlemissdata.com ________________ ► Subscribe to Our Channel Here: https://bit.ly/2TzgcOu Follow Us Here: Website: https://www.posit.co LinkedIn: https://www.linkedin.com/company/posit-software To join future data science hangouts, add to your calendar here: https://pos.it/dsh We'd love to have you join us in the conversation live! Thanks for hanging out with us!
image: thumbnail.jpg
Transcript#
This transcript was generated automatically and may contain errors.
Hi, everybody. Welcome back to the Data Science Hangout. I'm Rachel Dempsey. I lead customer marketing here at Posit. And I just want to add this in here because I learned that a few people are actually hearing about Posit for the first time through the Hangout. But I did just learn some people are hearing about Posit through the Hangout. So I want to add if Posit is new to you as well. We're the open source data science company building tools for the individual team and enterprise. So I'm so happy to have you joining us here today.
The Hangout is our open space to hear what's going on in the world of data across different industries, chat about data science leadership, and connect with others who are facing similar things as you. We get together here every Thursday at the same time, same place. So if you're watching this as a recording in the future on YouTube and want to join us live, there'll be details to add it to your calendar below.
But if this is your first Hangout today, a special welcome to you. Feel free to let us know in the chat as well so we can all say hi. We're all dedicated to keeping this a friendly and welcoming space for everyone and love to hear from you all no matter your years of experience, titles, languages that you work in, industries. There's also three ways you could jump in and ask questions or provide your own perspective. So you can raise your hand on Zoom and I will call on you to jump in. You can put questions in the Zoom chat and feel free to put a little star next to it if it's something you want me to read out loud instead. And then Curtis or KJ will share the Slido link in a second here where you can ask questions anonymously.
I'm so excited to be joined by my co-host for the day, Laura Ellis, VP of Data at Rapid7. And Laura, to kick us off here, would you be able to introduce yourself and share a little bit about your role, but also something you like to do outside of work too?
Okay. Thank you, Rachel, for having me on and reaching out. This is really exciting. So I'm VP of Data and AI at Rapid7. And what that means is I lead everything from our internal to external data and AI strategy. So that's things like our internal enterprise data warehouse, analytical tooling. I have some level of analysts that focus on internal analytics on my team. And then I also have data from a product perspective. So all of the products that we make, they obviously have data as well. And so our central data strategy there, and then leveraging that to do AI to impact our customer.
So usually, so I'm a mom, I have a husband, I have two girls, 11 and eight, and I have two dogs and we like to go hiking and stuff. And usually I'm like, I'm not that fun, but I actually am doing something kind of fun. So my 11 year old daughter is playing the guitar and she's pretty good at it. Actually, she's awesome. And she wanted me to learn as well with her. So since August, I've been learning how to play the guitar and she's so good. I actually really understand how good she is because she's so much better than me. And it's very difficult.
Data mishaps night
Laura, I also wanted to maybe kick this off with having you share a little bit about data mishaps night as well, because I learned recently that some people found about the hangout through a data mishaps night and conversation that was happening there. So I want to make sure we share a bit about it as well. Could you share a bit about what that event is and how the group came to be? Totally. Yeah. So data mishaps night, we just finished our fourth year. It's an event, virtual event that we run one time a year. My friend, Caitlin Huden and I, who actually met through Our Ladies, Posit, previously RStudio. And so we met through Our Ladies about eight years ago and then just became super good friends.
And it was during COVID that we would go on hikes together and we were just like, I miss the meetups. And so we decided to do this virtual event. But what we wanted to do is a little bit of a different take and talk about our mistakes. Because on LinkedIn, and I mean, me too, everyone's like, oh, all these great things we can do and everything we should do and all the successes. But you don't hear about the failures as much. And in data, because in data, data fails silently, right? So it can look real, but it's deceiving you. So it's really, really, really important that we're open to being wrong and we're open to asking questions and just sort of like being humble.
And so we wanted to have this event where everybody could share the data mistakes that they've made, get a little laugh out of it, learn a few things. And we just finished our fourth year and I think we got close to 400 people, which is amazing. If you think about that there's 400 people out there that want to spend an evening talking about data for two hours, it's pretty cool.
data fails silently, right? So it can look real, but it's deceiving you. So it's really, really, really important that we're open to being wrong and we're open to asking questions and just sort of like being humble.
Data culture at Rapid7
How do you carry that over into your role at Rapid7 and the data community within the company? Well, I probably should more formally carry it over. But personally, I think we have a pretty good culture of communicating and sharing and just being open about our mistakes and creating that safe space. I think that's what we're going for on the data teams as well, just because again, I think our successes and the things that we're good at are basically just a result of everything we've studied and all the failures we've made. So we really need to make sure that we're supporting that.
Laura, I know you gave us a little bit of an overview of all the different teams that kind of roll up to you as well and teams you work with. Would it be possible to share an example maybe of how you're using data across the organization just to put us all in your mindset and think of the types of business problems you're working on? Yeah, it's a good question because it honestly, it's really varied. It's really, really varied. So our internal analytical team focuses on everything and my enterprise data warehouse team, right? Like they focus on everything from running the business. So everything from running the business to creating like more strategic assets around value realization or even managing risk at the company, like risk of a variety of angles, even like product quality.
And then on the product side, we have a really cool few projects going on right now. So everybody's in this big race for AI, right? We all on this call know this and we too are running as fast as we can. Somebody said something like you're trying to outrun a tidal wave, basically. And we have some really cool AI projects. So we're in cyber, I'm in cybersecurity. It's just that, you know, Rapid7 is a cybersecurity company. And some of the biggest challenges with cybersecurity is it's such a volume game, right? Like in, if you're using our tools, you have to protect the entire landscape of all of your assets, all of your software, all of the granular pieces of software across on-prem to cloud and active threats.
And so people that are doing cybersecurity have a tough job. They're basically just looking at mountains of alerts and trying to decide like, where should I spend my time? Which vulnerability should I remediate? And so on the AI side, we leverage both the traditional ML to do things like kind of auto processing of alerts so that we can help people get to the things they need to look at faster. But we're also using LLMs to do, like, we have our own chat bot internally too. I think a lot of companies do, but we have our own proprietary one that our analysts use. We do auto generation of incident reports. The first draft with LLMs, we're using LLMs even to actually develop some of our vulnerability content in ways that would have been done completely different.
Making AI and analytics accessible
You have a passion for making AI and analytics accessible to everyone in a secure and scalable manner. And so I was just wondering, could you share a little bit more about how you're working towards that goal as well? So I think, because it's the two things, it's the secure and scalable and accessible. And so I think it's funny because in a lot of my journey, it would have been more analytically focused where, you know, we have to make sure that all data, this is not the fun stuff, but honestly, I kind of think it's fun. It's like, we have to categorize all of the data. We have to have policies. We need to make sure that all users have a role. And then we need to make sure that people have access granted as for the policies.
But I also think enablement is super important. So our team does every data set that we roll out, like we've gone through a huge transformation, every data set that you roll out, like we have to have documentation. Every new addition that we do, we usually have a blog on it. We have a monthly newsletter where it's like, if you're a data or AI nerd, you can see everything that's going on in the past month. Every tool has to have a getting started and a weekly office hours as well.
And then the AI side we're ramping up there, but it's pretty similar. Like we just got all of our policies finalized and our technical infrastructure to make sure we can see what models are out there. And then there's never enough AI people right now. Like there just aren't, there's not enough data scientists, all these wonderful people on this call, there aren't. So with these AWS services or Google services, how do we turn developers into data scientists, but keep them on the rails? So we do a lot of enablement as well.
That's actually something that came up last week as well, is talking about hiring people for these roles and how a lot of the job descriptions are asking for years of experience with AI when a lot of this is so new. And so I'm wondering, how do you think about hiring as well for those roles? Yeah, we have actually, right before this call, we were going over our road. I was meeting with our AI and data science team to go over our roadmap because, you know, we have sheets of projects that we want to do and not enough people. And so we're going through and saying like, what do we have to do ourselves and what can we teach people how to do? So we do like a research sprint with them and then we help bootstrap the making of it and then we teach them to run with it.
With all these cloud services where, no, we're not leveraging ChatGPT out of the box, but we do leverage like AWS Bedrock and like some of the cloud models, et cetera. You know, our developers already work in AWS. They can call those APIs and we can help them with the prompt engineering around that, for example. And so when we think about who to hire, I would say, yes, we do need more AI data science people who know what they're doing and can really focus on that research aspect. But we also just need to skill up a lot of the developers as well to be able to use the services and know the guardrails and know where and how to get help.
Data pipelines and the warehouse
I see Rogerio had a question in the chat. Can you talk a little about the data pipeline from the company databases to the traditional ML models? And there was a follow-on question, is the data warehouse still used?
So it's different for analytics versus production because I don't know, this is my biased view. Like, so Snowflake's obviously very, very popular for analytical workloads. I think it gets very expensive when, if you're talking about production, like product, data workloads. So they tend to be a little bit different, right? Like your analytical warehouse and pipelines versus the product. The analytical pipelines also, some of that can be batched. Like, it doesn't all have to be real time. You know, we can pay for tools that are a bit more focused on the end user versus, like the analytical user versus developer experience. And we don't have to pay for real time necessarily for everything on the analytic side.
Whereas on the product side, we probably need real time. We're probably looking more like Kafka and like Flink processing and more distributed query layer type of access, like Starburst, Trino, et cetera. And yeah, so I'd say that those are kind of, that's kind of what we're looking, we have more analytical patterns on the internal side, and then we have more of that like real time production level for product.
Data governance and access
I do see there's this kind of conversation that happens in the chat quite a bit about data stewardship, data governance. Shannon mentioned defining access and permissions can feel uphill, but once that's in place, everything works so much smoother. And so, I'd like to dive a little bit deeper into that topic as well and just understand what you're doing to help make that easier across the organization and defining access to certain data sets for everybody. Yeah. So, this is the interesting thing. So, in a past role, so I've always, I've been in data for almost 20 years, I've done all different data roles from like software engineering to databases, to data analysis, to data science, architecture, and then on to the leadership. And I kind of like dabbled in the security, but I never felt it so badly until we, in one of my previous roles, went to get FedRAMP certification. So, it's like federal compliance with your security protocols. And I can tell you that gets so serious so quickly.
And it was really hard, but it's almost like cathartic that once you kind of get it all organized, it feels so good. So, anyways, I say that because in a weird way, I actually really like this stuff. I really like the labeling and categorization of data and providing access. But the one thing I would say is, it's always better to keep it simple. It is always better to keep it simple. Less categories, less users, like less categories, data categories, keep them broad as you can. You'll have special exceptions, less roles of users. Again, you'll have some special exceptions, but try to keep the main categories for data pretty broad and the users for pretty broad so that the rules are pretty simple.
Because even though I think if you set out to it and you're asking everybody and you're trying to like make this beautiful thing, it's actually not practical and it's actually less secure because you have way too much to conform to. People don't understand the rules. They don't understand why they can't access the data. They start to develop these weird workarounds. It's better to keep it really simple, really transparent. That's the advice I would give. And then this is just one of those things that you just have to do it, but you can sleep really well when you do.
Career journey into data
I'd love to also talk a little bit more about your journey into data as well and a little bit about your career journey. So I'm wondering like was there maybe a turning point for you or when did you first like decide to go into the field of data? Yeah. So I did software engineering in school because this is why I love data though, actually. I bet everybody on this call has such a different path to data. It's so cool. I feel like there's not as many specialties where it's just such a wide range of people converging on data because it's like a language and everybody needs it. And then you kind of get that hooked and then you can invest more and more. So I came a bit more on the traditional path of software engineering.
I joined IBM one year out of school and I just happened to join the DB2 team, which is a database most people probably don't know on this call, but it was very popular back in the day. And that was my job. And I was like, I kind of like this. And then from there, dashboards, online dashboards became a new thing. This is, there was a time in the world when online dashboards were bleeding edge and I thought that is awesome. So I kind of, I followed that.
And then this is also kind of neat. Like I had my first child in 2012 and in Canada, no 2013, just so my daughter's born on New Year's so January 1st, 2013. And in Canada, you get one year off. So this was actually pretty cool because data this whole time and including until now has been on this upwards rise, like almost exponential. And so even back in 2012, the same thing was happening.
And I was just sort of like following it and seeing, and I was excited by it. But at the same time I was getting excited, the industry was too, and it was getting more accessible. But I went on maternity leave for one year. Before I went on maternity leave, I was doing analysis, but more like make me this report, make me this report. And nobody really cared about what I was doing because they just wanted their reports. I came back from maternity leave one year later and everyone was like, tell me about data. Tell me about data. Because there was this article that was posted, kind of a cheesy title, but it was like sexiest job of the 21st century. And everybody just got so excited by it. And they were stopping me to ask about my job. And I'm used to being the person sitting in a cubicle, just like pulling numbers.
And so then I ended up getting into data science and I went and got my master's because some of the questions that they were asking weren't just analysis. They weren't just like what happened. They were like, what will happen? Predictive analytics. How are these populations different? So you're seeing, you're doing like T-tests or whatever, ANOVA testing. And I was like, I don't know how to tell you these answers because we don't have data for that. I can't tell you what will happen in the future. What do you want from me? And that's when I started exploring statistics, getting into data science. Then I kind of went on a tear about realizing how horrible data actually is. And we should focus on the pipeline. That's how I got into the architecture data engineering side. And then you kind of have that thing where you're like, it's actually the people. And that's how I got into leadership. So that was my long-winded journey, but I just kind of kept following that next thing that seemed interesting.
Soft skills and leadership
I see there's a question from N. Bradley in the chat. Working in a field that puts so much pressure on hard skills, what soft skills did you feel benefited you most in your career to progress and through various leadership roles? Oh my gosh, right? It's so true because the thing is, and I feel like we've all heard the saying, like, what got you there? What got you here won't get you there. So you can build all your hard skills and then you get to promotion. And we do see it. Like, I do see it. Developers will get promoted to like lead or manager. And all of a sudden, they don't have any of those soft skills to actually get the job done, work with their peers, communicate what they're doing.
Hard skills, technical skills, it's super duper important, but also we have to be focusing on the soft skills too. And I think communication is like the number one thing. Communicating what you think they're asking you to do, what you're going to do, what you said you'd do, what you did, what you didn't do. You can't communicate too much. And then the second, maybe just off the top of my head is probably partnership. I think a lot of on the technical side, we're really incentivized to produce, produce, produce, produce, produce. But once you get into leadership, your version of producing is really like through influence, through others, like with influence, with others, whether it's your team or across or up. And so you really have to have that real partnership focus to be able to drive the results.
Data security and AI governance
Curtis, I think you had a question a little bit earlier. I think it was on maybe data security issues. Do you want to ask that question? Yeah, I mean, sure. Since you have quite a bit of experience with data security, and there's a lot of practitioners and leaders in here, I was kind of curious if you had maybe examples or advice on the kind of the big vectors of concern as an individual contributor or an org should be worried about.
So I think, again, it depends on the systems. But at the end of the day, when you're thinking about data security, you have to, I feel like it's so simple, so I hope it's even helpful, but you have to be able to see everything, first of all. So you have to get it all down, get it all visible. What are all of the systems, let me say this, which store or process your data? Because they're all in scope. So identify all of the systems which store or process your data. And then you have to identify what is the data. And that kind of goes to the categories thing, but you have to really get the lay of the land. And then you have to identify, what are we allowed to do with it? But not just as per our opinion, as per our legal policies, and then our company values and ethics. And those are the most important pieces.
See everything, understand what it is, and how can we use it. And then it's really like a matching game therein. And then you make the policies, you implement the policies, and you monitor and audit the policies. It sounds so boring, but it's really, it's that simple. And it can seem really daunting, right? It can seem really daunting. Everyone's coming up to you with their own special thing, but you just have to stay focused on doing those core buckets, setting up your foundation, and then you can build from there.
And the reason I'll say this, the reason I'm trying to be like, it's not that hard. I think everybody, I think people in the industry have worked out their basic data governance. You just got to work through it. But I think, I just released a blog on this. We're running, we are running towards a disaster on the AI governance front. It's the same thing. You got to see what you're working with. You got to know the data it feeds in. You got to know how you're allowed to use it. You got to have a policy. You got to implement. You got to audit. And people are making AI governance into this scary, unachievable thing. And if we thought data governance was a problem in the past, AI governance is going to be our nightmare in five years if we don't get a handle on it.
We're running, we are running towards a disaster on the AI governance front. It's the same thing. You got to see what you're working with. You got to know the data it feeds in. You got to know how you're allowed to use it. You got to have a policy. You got to implement. You got to audit. And people are making AI governance into this scary, unachievable thing. And if we thought data governance was a problem in the past, AI governance is going to be our nightmare in five years if we don't get a handle on it.
No, I was just going to say, I love the way that you make everything sound so practical. Don't make it so scary. This is how you should do it. But then I was like, oh, it got a little scary towards the end. Yeah. So yes, you're right. It is not scary. None of it's scary. And I'm like, just do the same thing for AI. It's also not scary. What's scary is if you don't do that. So just make your basic foundation layer. Because people will say, I'll talk to them like, I only have two models. Why do I need to do that? I'm like, yes, because you only have two models now. Guess what? Your company is going out making all these models. Why wait?
AI as a profession and threat detection
There was an anonymous question, should we consider AI to be something that current analysts can pick up as part of their job? Or should we think of it as a new profession? I think, so it's a hard one. I think that it's, it's emerging as a different set of, it depends how you call yourself, how you define an analyst, because honestly, that is hard too. We have like data analysts, business analysts, we have analytic engineers, we have data engineers. So I think we have this like spectrum anyways, with a whole lot of overlap.
So I do think AI warrants its own title. Now it may have a 90% overlap with what a data scientist is. And a data scientist may have an 80% overlap with an analyst. So I wouldn't, I wouldn't get caught up on the roles. I think it's enough of a new area that it does warrant at least making room for a title like that, because we're dealing with concepts that we didn't really have in broad population before, like LLMs. And they have different considerations around them, around data leakage. They have sort of different out-of-the-box, like really viable out-of-the-box cloud services that we didn't have before.
LLMs really do leak data. LLMs really do say crazy things. And they really do not always have transparency. And even when they have transparency, sometimes it's not real. So I think it warrants having a title. Whether an analyst also can do that, I think go for it. I think anybody can do that. I think AI and a lot of what I just described is super accessible. And that's why everyone's talking about it so much, because it's become so accessible to the general population. So I'd say, regardless of your title, go for it if you're interested. And probably every single profession out there will be using some level of AI in the next five years or two years even.
I see Brian had a question. Are there any special strategies or difficulties with using data to identify threat actors or attack patterns? Yes. So there's definitely lots you can do out there, because all of our computers are generating oodles of data which encode kind of what's going on. And we do already at the company leverage a lot of machine learning to be able to identify patterns, whether it's patterns of type of attacks. So if we're looking for a particular type of attack, you know, we can look through all the logs and then start to say when this, this, this, this, this happens, it's typically this type of attack.
You can also, so that's more like on the detection side. I think ML does a really good job being able to detect threats, active threats, on the what do I do about this, which is like alert triage. So you get a detection, it pops up alert, boom, boom, boom, look at this thing, right? Well, of course, you want as many of those as possible, but you're going to get so many. So how do you prioritize or get rid of the volume? Also, ML is specifically pretty good at that as well, being able to do things like clustering. So these type of alerts are typically, they all are part of the same attack. So rather than an analyst, just looking through them and having to make the association themselves, we can look at prior patterns and put clusters together. So they should look at them all. And then, you know, relative to a customer's environment or their baseline patterns, we could say, oh, actually, this is a, this is more malicious than we would think, because this customer never does this type of thing, right?
So we can boost up the signal or reduce the signal. And then on the response side as well, that's where you can start really, that's where like, so we already did those two things, but that's where you can really start leveraging the LLM side of the house, because you can start baking in context into the investigation that they're doing from other data and summarize it with the LLM. So analysts can get data easier, easier, easier.
So I guess the answer is like, yes, there's so much you can do because cybersecurity is all about digital trails, which is just data at the end of the day. The hard thing, limitations, the hard thing is that A, attackers are changing their patterns all the time. So where you think machine learning AI is basically trying to predict based on past, but they're building new patterns every day. So that is difficult. We can do anomaly baselines and things. And then the second thing that is hard too, is that they're using AI as well. So as we're beefing up with AI, they're using it too, because they're also essentially a resource constrained entity too, that is leveraging AI to also go better, faster, stronger.
Career advice and getting started
Doug just asked the question in the chat. Seems it's very hard unless you have three to five years of solid data science experience to get a first data science job or first ML job in particular. Yeah, I would say, well, there's a kind of cool thing that's happening now too though is, so there's the postings for data science and AI and ML. But I would suggest just getting like a data type of job and then starting to deliver AI and ML. If you, if you can't, if you want to do the data science, AI, ML, whatever we call it, if you want to do a job like that and you're finding it too hard to get it, my, my thing is get your foot in the door, get a job doing some level of a data job, and then just start delivering as well. I mean, you're going to have to put in some of your own time because you got to do the job you have as well, but I, I would start finding opportunities and work to deliver that, and then you can kind of like transition your career as you go.
Yeah, I think one of the things that I would really recommend strongly for folks is, and I learned this myself when I'm hiring, is I'm looking for folks who are producing some level of work. So being able to share your work, whether you're just starting to produce, you know, work in a GitHub account, that could be something as simple as taking a data set, exploring it, building visualizations, and going through a traditional VDA workflow, evaluating and deriving inferences from that, and just as you would in an ordinary job where you're giving takeaways or recommendations to the stakeholder. I'd also recommend if folks invest small amounts of time in just writing up small documents. One of the things that people use LinkedIn all the time for is to generate ideas. And they're looking for ideas from people who can engage with data and communicate. I think that as a data science leader, I think I've spent a fair amount of time, half and half, between folks who prefer working with machines and folks who prefer working with people. And so bridging that gap and being able to communicate really sets yourself apart from probably a dominant majority of folks. And that is a great opportunity to show your work. So I would say, you know, publish if you can, small amounts of time will pay you big dividends.
Hi, Laura. Long time. I think definitely putting yourself out there, right? Add into what Peter said about publishing, whether that's an article on LinkedIn or just like, you know, a quick TikTok reel about, hey, this is the school data set I found. These were the findings. And, you know, here's a link to whatever, you know, artifact you have. I think also looking at meetups, community meetups, they are always on the hunt for speakers. I know it can be scary to speak publicly, but, you know, just try to give it a shot, see what it feels like. Because what that happens is that eventually people start thinking of you as an expert when you're out there on the stage talking about different things. And you gain that confidence and it also increases your visibility to potential employers, to potential managers. So I think combining that with the networking aspect of it is another way to, you know, stand out a bit.
I couldn't agree more with both of those. Those are so good too, right? The learning out loud. Absolutely. Well, I know as we start to get towards the end of time, I want to make sure I always ask this question. But as you look back and think about career advice that you've either received over your career or given to others on your team, is there a piece of career advice you'd like to share with us?
So I think there's two. So the one I always say is like, just get started, whatever it is. Whether it's that or whether it's learning a new AIML technique or, you know, getting, learning dbt, I'm just making it up, right? Like, just get started. You just keep going. And then it's so funny when you just keep going and you just take all those gremlins off your shoulder that are telling you, you can't, you're not good enough. Who do you think you are? Right? Like, just get them off, get going. And, you know, to Grishna's point, what you'll do is you'll find out like five months, one year later, two years later, all of a sudden, you're actually, you've built some skills. You're pretty good at this. Maybe you're even seen as an expert, right?
So the just get started is my number one. And then the other thing is, it's like, you know, I don't know about the people in this room, but sometimes you can find yourself in a position where you've been given a big project or something and you think, and especially like people in this, I mean, data and AI, data science, whatever that we call it is changing so quickly. You know, I think probably a lot of people in this room are getting these new challenging projects or growing themselves in one way or another. And you're thinking like, can I do this? Should this really be me? Like, shouldn't, is there somebody more qualified than me to be doing this or speaking about this or whatever? And I would just say like, you know, everybody here is entirely unique, brings entirely their own set of skills to the table, their own perspective, life experiences, whether it's, you know, all these hard skills or all these tech skills, or, you know, more breadth of leadership or responsibility experience. There's nobody that has a super set of your skills out there. There's literally nobody who has lived your life and then has everything else you don't have. So trust yourself that you're pursuing this thing, you're in that new role, you're having that new project for a reason, and you will be good.
I would say, too, like that kind of speaks to Peter and Grishma, especially his point where it's like, yeah, you have all these floods in, but we also know people in communities, communities like this, you know, that's where it's great to go out there and make those connections, even if you've never done an ML job, AI job, do science, go reach out to people on LinkedIn. Like, hey, I would love to do something like you in the future. I would love to work on your team in the future. Can I ask you a bit about what your daily life looks like? And then I know you. I think a lot. I think, you know, Benjamin and Doug, probably you would agree. A lot of hiring is also done through people that, you know, as well.
Laura, I I'm so impressed by all that you do in the community and seeing data mishaps and I reading your blog posts and I'm wondering how do you set aside time to keep up with that part of it as well? And if you have any advice for us on doing the same. Thank you. That was very nice. I mean, honestly, I'm I'm like stressed out. I'm always running around trying to do things, but I really like it. That's that's really honestly the thing is like I wouldn't do it if I didn't like it. The data mishaps like Caitlin and I are are. I mean, it's like you could be on a work trip and you're sending out the link to the Zoom chat or it's it feels a little bit. I mean, honestly, sometimes I'm like, why did I do that? Why am I doing this? Even I just released that AI governance blog and I released it and I thought to myself, does anybody even care about AI governance? You know, but I do. I care about it. And so I wanted to get my thoughts out there and the data mishaps night it like it fills our bucket. You know, it is hard. I wouldn't say I have any at all secret for success to doing it all except for that. I really like it and care. And I think people innately make room for the things that they really do like and that fill our bucket.
That's great. Well, thank you. I'm glad that that is in your filler bucket and we all get the benefit from that in the community here. But thank you so much, Laura, for taking the time to join us. I really appreciate it. Thank you all for the great questions. I love getting to go back and reread the chat after because I don't get to do it at the same time. But have a great rest of the day. Everybody. Thank you. Have a great day.