GitHub Open Source Friday, 4 Aug 2023

Previous Meeting Next Meeting

⏯

youtube image

►

From YouTube: Bridging Open Source with Actuarial Science

Description

Open source is eating the world. It's even impacting our daily professional endeavors Today at 1pm ET, we're chatting with Matthew Caseres about building a community of actuaries and actuarial software in the open.

A

A

A

A

A

Hello- everyone, it's nice to see you all again. um If this is your first time watching, this is open source Friday, it's a weekly, Friday's, twitch stream that I do or I guess multiple platform stream that I do um basically chatting with open source, maintainers and core contributors about the work that they're doing in open source. This is to help us get a better understanding like us, as viewers get a better understanding of how we can get involved in open source and also help us learn more about other open source communities that are out there.

A

um I see people in the comments saying hi, someone said: hi, GitHub, family, hey Naveen! What's up, um let us know where y'all are tuning in from someone said they're tuning in from Italy um and I. Don't know how to even I don't want to like mess this up, but thank you for tuning in from from that country, um so I have without further Ado instead of me, just talking on and on I do want us to take some time to chat with our guest Matthew, hey Matthew. What's going on.

B

I'm, just sitting down ready to talk to you, I guess about open source Actuarial software.

A

Yeah I love, it love it. uh Could you tell us.

A

Just so people get an idea of your background and who you are.

B

Yeah so I started as an actuary after I got a math degree. I went to University of Alabama at Birmingham, got a math degree, wasn't sure what to do with it took some Actuarial exams became an actuary and started writing code, and the department I was in wasn't really code writing department, but it was a good solution for the problem. It didn't fit well within Excel, which is where actuaries do a lot of work. So I started writing our programming scripts um and eventually ended up publishing my first R programming package.

B

um Maybe you know about a year after I started that job, and that was my start of Open Source Actuarial stuff, and that was about four years ago and um so basically, at a certain point, I started having some health issues, I guess you'd, say uh just not taking good care of myself and ended up taking a break from working for a while and when I wasn't working.

B

I stayed engaged by writing open source code by learning new programming languages, so I was just an R programmer and I started doing a lot of typescript and a lot of python, and so that typescript knowledge allowed me to start building websites for the open source community and to really just have more opportunities in general.

B

You know um so that is kind of so I took a couple years off and I wasn't really even thinking about Actuarial stuff too much, and then some people reached out to me based on work that I had already been doing, and they said, hey look at this guy. He used to do this stuff, let's add them to our LinkedIn group.

B

So someone made a LinkedIn group and they found me because they knew about me and they added me to the LinkedIn group and I ended up building a website for their LinkedIn group and now I get to participate in this uh small community of Open, Source, actuaries and I love, looking at all the different projects and talking to the maintainers and seeing what they think is interesting and and so it's just something. I have a lot of fun doing now.

A

I love it that I love that it's such a realistic story like you started off in a job, you became a career changer due to like health and stuff. Like that things like that happen, and then it sounds like you know, you spent that time.

A

um Just learning, learning about code, more about coding, learning more about Open Source and getting involved in different communities. Okay, I have a really embarrassing question. First off Laura, hey Laura, um she said: oh, it's team math day. I purposely didn't ask you this. When I first met you and then like, even when we were backstage, but what is it actuary? Because in my head, I thought it was an accountant, but then, whenever I hear you talk, you're like talking about our programming and stuff like they combed.

B

Right and and so basically, an insurance company has to pay money for claims and a claim, and at least in life insurance would be if somebody has a life insurance policy and then they die. That would be a claim, and so we don't know when people are going to die and that's the most common actuary joke is. Oh, you can tell me when I die. Well, they don't really know when you're going to die, but there's some future uncertainty around.

B

What's going to happen with the insurance claims, and so the actuaries are responsible for maintaining the financial health of the company and to estimate future risks. So an accountant deals with cash flows that exist and are certain maybe and the actuary is like an accountant but maybe dealing with cash flows that are uncertain and there's people that know a lot more than I do I was in the industry for about two years and a kind of specialized role, yeah.

A

Gotcha and you all use code to help you make those predictions instead.

B

Well, sometimes we don't all use code, but it's it's it's it's interesting. You know is Excel code Excel. You know it has formulas.

A

B

Does stuff and it has VBA in it right and maybe we have an access database or something.

A

B

There's different levels of kind of proprietariness and standardness of a technology stack and so actuaries are beginning to adopt more you'd, say, standard technology Stacks, but there are still proprietary, Computing platforms that are, you know, really uh integrated in with a lot of major companies, and so it's not clear where the pendulum will swing in terms of open platforms and closed platforms and how that's all going to pan out over the long term. Awesome.

A

Okay, that makes a lot of sense. Thank you for that clarification um really quickly. I'll, say hi to a couple more people and then we'll get into like what exactly is your project and a little bit more about like what inspired you to get involved? You kind of gave this a little bit but just curious. um So we have someone joining in from Amsterdam. They said they're, another actuary, actually, that's exciting, um someone from Zambia Vietnam, Ukraine India, so we're all over the world.

A

All Global I love to see that someone from Pakistan and um Madagascar in Dubai. It's so crazy! How you all like find this but I guess because it's GitHub like it's. A lot of people are using GitHub um and thank you all for tuning in also I, see learner0418 they tune in every week. So really excited about that. um Thank you for tuning in all right. So what exactly is your project um that you're coming here to talk about sure.

B

I mean I, run a website um and I guess you could share the screen if you want, if you're able to do that, um so I run this website and we have some fun features like internationalized. It's kind of a gimmick but I think it's fun, and this is why Actuarial open source and one thing is the Actuarial science there's a lot of symbols.

B

There's a lot of special knowledge and if we're able to turn all of that into code, then people are going to maybe be able to read it a little better and then it's not just something that exists in a book. You can, you know, press a button or run a terminal command and you can run it and so I think that code is really valuable for in an Actuarial context, because we can move from having theories and symbols to having things that actually run and do work. So I'm excited about communicating Actuarial Concepts through code.

B

um We also have open source allows for reproducible, research and Actuarial science. A lot of these data sets are closed. Data sets it's some proprietary Insurance data set they're, going to publish a report on something and it's going to be published as a PDF I think that we can have a lot more value in people's reports and in people's research. If they start having open data sets, maybe they anonymize it.

B

Maybe they have a data simulator and then they're going to have reproducible research that end to end everything can be replicated, and that's really not the case in a lot of Industry Publications, currently, and and that's why we have the open data and non-proprietary Computing platforms is just a lot of the computation. That's done exists within closed Computing platforms. You could think like Matlab has certain closed platforms for engineers to do their work, so it's kind of like special Actuarial software, which you would need a license to use and there's not any information about it.

B

That's generally available, so reproducible research is exciting and also anyone can get involved. Collaborative space professionals, academics and enthusiasts I. Don't currently have any credentials or letters next to my name or anything like that, so I think I'd be an Enthusiast right now, but really anybody can kind of start looking around and getting involved and seeing you know if there's something exciting or something they want to.

B

You know get involved with, and then we have this project display, and this is just some projects and if you click this button, then you can see all of the projects and there's more projects that I probably need to add and uh I Implement. These little features- and you know I'm sure I- could make this better, but uh we have these projects and then we have some motivational quotes up there in German, okay, we have some motivational quotes and I really like these quotes and I just try and keep them.

B

In my mind when I do things like open source is about collaborating, not competing, and uh it's just it's just fun. It gets me excited yeah,.

A

No, this is exciting because I do think you know.

A

Of course we are in the know about open source and the value of it, but I think it's really interesting and exciting that you're bringing it to things like um like Actuarial science and all of that, because I, don't I mean maybe I'm, just not like super into in the know of that area, but I do think other areas or other fields can really benefit from from open source and like using open source principles like you're, saying, like making things more accessible, making people be able to run the code or run these formulas um inside of a terminal, rather than just looking at like symbols and stuff like that, just making it more accessible for people, so I do think this is exciting and I like that.

A

Y'all um applied applied this to that. We also have somebody saying or learner0418 saying very exciting, to see open research and collaboration and then I guess someone's remembering a movie called The actuary I've, never I've never watched it, but.

B

Just so can I show a couple more features: I actually have a couple things, I'm excited and so I might go through them quickly and then, if there's something you want me to expand on. So this feature isn't quite I, guess you'd, say shipped because it's not accessible from here kind of because we committed to this internationalization thing and it would be very hard to internationalize a lot of text.

B

But we have these interactive python exercises I kind of put this editor together myself with some open source components and we can run Python and the idea is maybe we could do some educational stuff, I really like interactive educational exercises, I love leak, code and stuff like that, and so maybe we could have some exercises where you can grade them and they can help. You learn about Actuarial science with python and we have just some stuff.

B

You know we have some stuff there's some exercises, but I would really like to implement grading like there's this website I just kind of work on it sometimes night, sometimes weekends and sometimes I- take a break from it.

B

So there's a lot of architectural improvements that you know, I might need to make to this website to make it stable and then just one more thing is that there is this I just think is cute is that you can have a react component and then you can type type some stuff and then I do this I press this button and now I actually have the jsx. Oh.

A

That's so cool the.

B

Component itself, so I like and so there's a toggle there's a prop that you'd pass in. That would say this component is an in editor mode, and then you would do that and then you click this button. And now you have the jsx for a component which is not an editor mode but has the same content so I think it's a fun kind of content, authoring workflow that.

A

Is so cool I actually love this, and when you said you wanted to implement grading, did you mean like I, wanted to make sure I understood you did you mean like? um Did you mean, like? Oh? Did the person put in the right like code.

B

Yeah, okay, really really! You know there are test cases and it's kind of leak code style, but this is running in webassembly. It's not connected to a server, so I don't have to run some server.

B

That has some kind of sandboxed python that is executing this code, so using webassembly to execute python code and having a nice way to grade the questions and then a way for people to look through and have the questions that they want to and like just uh how to say a way to navigate through a collection of questions where there's progress, indicators and stuff like that really trying to build and education platform, that's Pleasant to use- and you know maybe it's just for Actuarial stuff over here.

B

But if people want to do something else with it, then it could be a nice template.

B

So we're really not there I think that we want to migrate to um the app router features of nexjs 13.4 and uh right now we haven't and I'm kind of waiting for a big chunk of time, I'm in school, so I don't have time to ship these big changes so maybe someday we'll have app router, which will make it easier to do these kind of server side things and you know, have optimistic, rendering and stuff like that, and then the educational platform would probably be a little bit.

B

It'd be it'd, be ready, so I'll finish school in maybe a year, and then maybe I'll do it or maybe I won't because there's other stuff that I'm excited about like the actual kind of um kind of computational research, open, uh reproducible research- so here is some kind of re reproducible. Research would be in this benchmarks repository and this benchmarks repository is something that we run and we pay someone um I, guess fifty dollars an hour for four hours a week to try and ship improvements to it and that's funded through our open Collective.

B

So we have this open. Collective I put about five thousand dollars. My savings into it a while ago, so our current burn rate is ten thousand dollars a year and we maintain these benchmarks where we have various Actuarial tasks. How long do they take to run and do they all produce the same result? So this is something that's a lot of fun and we want to publish this to journal of Open Source software or something like that.

A

That's so amazing, first off that you believed in the the project so much that you like invested your own money, but this is so much cool stuff that you're doing even the the educational side. It's almost like you're doing like a or hoping to do like a code academy for actuaries I.

B

Love it and it's a lot of complexity, so it just depends on you know. Maybe I will never finish it. If I don't get the passion to do it or maybe I will so it. It really just depends on. You know what happens with my life, but regardless of where things go, I think it's a lot of fun and it's a lot of learning that I get to do as well. This.

A

Is so fun and I? Think okay, I have like a limited understanding of like the open research part of the this like? Can you explain more on like how this is beneficial? I do see actuaries like saying already they're, like all their actuary professionals, will benefit a lot by having a place to check practical examples, great idea, but yeah can it help me understand this part a little bit more yeah.

B

Yeah, so this is various calculations, and so I used to be a life actuary I'm, currently in more of a car insurance type company, not exactly more of a commercial lines, you'd say, but so my knowledge is mostly about life. Actuaries and uh modeling life insurance is kind of its own discipline, so we have something which would be a basic term. Life insurance policy, a basic term life insurance policy means that you know if you die within the next 10 years. You'll get the payout.

B

So this is an example of a life insurance model which comes from a project from a really great actuary. Well, I think model X, okay, I I went to the long, the wrong link life web okay. So this is built by the person who made the LinkedIn group. So by the way we have a LinkedIn group, LinkedIn group, Actuarial, opensource.org join the LinkedIn group, I, don't know, and we also have a GitHub organization. Both of these were set up by a Japanese actuary at PWC.

B

Really recognized this kind of problem that I'm interested in a while ago and about 2018 I believe he started working on life flip, and this is really a full implementation of life, insurance, Actuarial, modeling software, and he has used his own kind of tech stack to make a number of libraries that talk about various life insurance problems.

A

B

Life and so I, you know this is Actuarial modeling software and my interest has typically been more in how do I use Python by itself or just a program Language by itself, and so this is a little heavy, but it does implement features that actuaries are very used to and that they expect from a solution for life insurance modeling.

B

There is a graphical user interface that is implemented in the spider IDE for Python and I'm, more of a vs code user, so I want something that works in vs code, but maybe he'll come out with something. But so these are a lot of life insurance models that he has built, and so what I did is I started the benchmarks, Repository story, and so here we have python Lively basic term M.

B

We see that it takes 689 milliseconds for our sample calculation, these run in GitHub actions and automatically populate everything they do they do, and so this is python Lively basic term M I went and did a implementation from scratch, which is producing the same result, which is 553 and then I did an implementation. That is more of an array-based programming. These programming methods are recursive and use. Caching, it's similar to top-down dynamic programming, if you're familiar with that, but this Jax is basically array based.

B

You could do it in pi, torch or just numpy, and this is some array-based kind of way of doing the same calculation, and this is just a Julia implementation that we have and we're working on, more Julia, implementations and then yeah. So.

A

B

Another life insurance model- this is something that is just about mortality tables. So we have different mortality table software and we made a little Benchmark for those mortality table software and then this is something saying just another thing. So we have these tasks and I just went ahead and set them all up when we set these up.

B

Originally, we saw that there was a discrepancy between the number of rows that they produced and so I went and opened issues on the relevant repositories and worked to close them, at least on the Julia side, the our guy fixed them right away pretty quickly, and then I went and kind of reworked a lot of stuff within this repository and uh submitted a pull request so that we can have these numbers sit here and be the same.

A

This is so cool, like my mind, is blown I didn't know, that's. This is how like insurance and stuff worked at all, so this is so cool. Wait, uh I have some questions. There's questions in the the the chat that I'm starring, but I do want to just quickly see. Are you able to show the GitHub action by any chance, because now I'm curious about it? Well,.

B

Sure yeah, so, let's go to yeah, so I'm, just going to press dot, I suppose yeah I love this feature: I love, GitHub, so anyways.

B

um So we have this file generate readme.pi, which is going through so each of these um each of these, so we have Julia, Python and R Julia, Python and R. Each are going to execute some set of benchmarks and generate a yaml file. Now this yaml file isn't perfect, because then we would have how to say we want things readme.md. We want things to be grouped by the type of Benchmark. We.

A

Don't want to be.

B

Grouped by the language, so we have to do some processing on these yaml files bills wherever they are. We have to do some processing on these yaml files, so that things are grouped by this, so to kind of get the format incorrect. We have generate readme.pi which goes through and does some sort of thing. Who knows what this does right.

A

But it goes through it.

B

Does some things to do the formatting and then it is going to readme.write template what is template? Let me just to control. Okay: template equals F, dot, read readme.template.md, so this is the template right.

A

B

Some template and then this readme generator readme.pi is going to take the template and then it's just going to concatenate the results below it. Oh, so that's what generate readme.pi does. If we look into the GitHub actions, then basically it sets up our. It sets up Julia. It sets up python. Now, if we look into python, say we're going to check out set up python, pip install and then we have this upload artifact step and we upload these artifacts and then to create the readme.

B

We depend on benjar bench, Julia, bench Python and then we're going to download the artifacts from the previous steps so that they're all accessible. Oh.

A

B

Yeah, so it was a fun little bit of work and just one thing that I really like about uh GitHub actions is just the it's very common. So if you start asking gpt4 questions about a GitHub action, it will just totally tear through that, for you and uh I hear there's some new tooling about like a GitHub actions, um kind of extension for vs code or.

A

B

A

B

Github actions it's great to work within um kind of a tooling that is so standard and so widely used it just it's great yeah.

A

No I love it. It's what it's! What got me interested I GitHub before I work there so love it wait. Thank you for showing us that there's a question here from um someone a viewer on YouTube, they said awesome. Do you also touch on competing risk events.

B

I, don't even know what that is to be honest, so I I really do have a pretty narrow scope of understanding. I was in the industry for two years and I was doing data analysis to support um assumptions that go into the model. So I would do some data analysis and I would say. Oh, it looks like one percent of people die if they're in this demographic or something like that and that's a simplification.

B

So that was what I did professionally and since leaving the industry I've been relying entirely on open source work from other people to kind of continue growing. My Actuarial knowledge gotcha.

A

Gotcha, that's that makes a lot of sense uh and then someone else just made a comment saying many big companies might be interested in supporting projects like this Tesla with car insurance, Uber, Insurance drivers and careers Etc well,.

B

Yeah my thoughts on that are we're. Gonna go GitHub, casualty, Actuarial, okay, here we are so. This is a professional Society. An Actuarial life is kind of dominated by professional societies who give you credentials after after you take some series of exams and these credentials allow you to sign off on regulatory paperwork, so casualty.

B

Actuarial Society has a GitHub account and I think it's the coolest thing ever so what this person was talking about, he's saying uh Tesla with a car, insurance and Uber is also a car, so uh both of those would be uh typically uh kind of people involved in those activities would be members of the casualty, Actuarial, Society, gotcha and so I have all my experience in life. Insurance I couldn't say too much about it, but this is actually the most popular piece of Insurance Open Source software in the world, and it only has 145 Stars.

B

So you guys should all smash the star button.

A

B

But this is, uh this is just really cool and it's got a great documentation site, and uh this is kind of what would be relevant. I think that if Tesla was interested in supporting someone that uh it would make a lot of sense to support the casualty Actuarial society's efforts, because they just you know they have resources, they have people and I. Think that they're doing a great job.

A

I'm trying to find it a team, lighter python, I'm, put it up on this folks, um okay, this is this is really cool like this.

B

A

A whole nother world that I'm getting exposed to um there is somebody in the chat, and this is a question I was going to ask anyways- is how can people get involved in this.

B

It's um you know it's hard to say because it's like I'm not having the most time to try to help someone help me if that makes any sense. Sometimes I get an open issue. Someone wants to implement a feature and I just say: I'm, not sure I want the feature, and even if I did I'm, not sure I'd have time. So, if somebody, if I, really trusted them, then I would just let them do whatever they wanted like.

B

If I looked at their repositories like if I knew who they were, and they were then I would just be like yeah just take control of the website or something.

B

But it's not really like I know anyone so the benchmarks with the benchmarks, the uh the guy that um I pay has he he just has whatever permissions he needs to do whatever he wants, um but I think that you know getting involved in Actuarial, open source, I I, don't know how easy it would be because it's pretty specialized and you might not know what people want so overall I would just encourage people to get involved in open source and to just contribute wherever they can and to use their own knowledge and their own needs and to just what are they passionate about really so I I just kind of stumbled into this, because I have some experience and I've seen this stuff I mean I'll, be basically if you go and you start reading through these packages.

B

If you start understanding, Actuarial science, you start working towards maybe even becoming an actuary, then in the long term, there's a lot of stuff that you can do. But overall, it's not so clear to me that there's an easy path to contributing, because you don't know what the needs of business are and you don't necess yeah, really that I guess that's.

A

That's theater I, wonder if there's like, maybe they can help in terms of like project management or something like.

B

That and there are open issues, and some of these issues are administrative in a way you.

A

B

Find a project that's not tested, you could find a project that has something wrong with it. You know, but um but yeah. So you can just look at different projects, but I guess I I, don't really know how to contribute. I have these benchmarks that I work on and I have this website that I work on.

B

If somebody is an absolute beast at xjs, 13.4 app router, please run the website, but uh please have some high quality work so that I can trust you and then maybe you can get some funding from the open Collective, but it wouldn't be that much. It would not be as much as we pay our other friend just because I do not have that much money.

A

Yeah and what wait I have a I, have a selfish GitHub question or a promo question: I! Guess right now, if you have sponsors, why.

B

Don't you have sponsors well at the time I just wasn't aware of it. I mean I started this a while ago and I started it, because um I had seen an open Collective for unified JS, which is some markdown parsing software, which I used to do a lot when I was in more educational technology interested, um but I do actually sponsor someone on GitHub.

B

So if we uh my profile I sponsor Ellie, Lam NZ, uh who actually makes uh some the kind of the Python running component on the website, oh that's cool because it just saved me so much time yeah. It was like. Well, if I give him five bucks a month, maybe I can do that for a year and if I cancel or change to someone else, but you know they worked on it a bunch and it helped me out a bunch and I spend money on stupid stuff.

B

All the time so I might as well spend money on someone that did some really great work. Yeah.

A

No, that's the way to think about it, especially if you're leveraging their tool, and it helps to make your your life a little bit easier.

B

Yeah and so I also opened a bunch of issues on their stuff, so maybe they'll change it, but I. Just I didn't I I'm, actually uh kind of just taking the internals from their software and reworking it as necessary. um But still it's such a solid foundation and and yeah.

A

Any okay, so any advice for if okay, they can't get involved in well, not they can't, but it might be harder to get involved in actual real open source, but any other things that helped you to get involved in open source in general and.

B

Actually, one thing I want to say is that the benchmarks- you know they're pretty concrete um stuff, so if you actually want to make something faster and you think you can make it faster, maybe I misspoke earlier, if you can make some of these benchmarks run faster and you want to go for it, uh go for it and submit a PR really um and also like there's a lot of computations uh from this uh okay.

B

If there's a lot of computations in life lib in life lib, we only have two benchmarks from lifelib, but maybe there's more things in life. I mean I, know, there's a lot more stuff in life lib.

B

So it's just a question of if you know if we want to have those as benchmarks, if if we should have those or not and and yeah, and also if someone wants to uh give us money, you know but I'm about to re I'm about to refill the funds so I put about 5 000 in uh or down and I'm, probably gonna bring it back up to 5000 within the next couple of weeks. Yeah.

A

Yeah, that's a good point, there's other ways to contribute in terms of like financially in terms of just using the tool testing it there's there's lots of different ways, um but I kind of want to Pivot into like just your open source experience a little bit just talking about that. So um as a as a and I think you kind of went into this as like a core contributor or someone maintaining a website. What are some challenges that you might have like come across or any struggles that you've had um just navigating all this.

B

I think the kind of part-time nature of it, and also that you're responsible for everything, there's no Tech lead to tell you hey, that's nonsense, don't do it like I can do whatever I want, like I just tried to push a hotfix to something I just tried to do a ninja edit without even running the software, like literally from github.dev, without running the website and verifying that it even worked, and it failed.

B

So it's like the part-time nature of it is challenging being a student while working full-time, wow, sometimes trying to train for like road races or trying to be involved in some Endurance Sports I mean it's yeah.

A

B

Seven hours a week of training, but it's it's you know stuff doesn't get done. Stuff gets done. The very slow pace but I'm. Okay, with that yeah.

A

Yeah you're doing a lot of different things and then what's what's something that's been rewarding to you um in this space.

B

um I, just like I I guess.

B

I guess getting to kind of have working relationships with different maintainers is something that I really like, and you know I think when you get involved in this stuff, you might think. Oh, maybe if I do a bunch of stuff I can get on open source Friday and that's what might motivate people the maybe someday the flashing lights and I'll be a star I'll, be a celebrity of Open Source or you know whatever might motivate someone to do the work and after you do it for a while. You start to be like.

B

Oh, this guy asked me to do a code review for him and he's writing cool software and I get to see this code review and think about it or I get to open this issue I'm using this stuff. Just having like a kind of this working relationship and being able to contribute to other people's stuff and other people contributing to your stuff, you know: we've got three: we've got two people that are not me that did contributions to the benchmarks, repository for example, so.

A

B

Love I I, just like seeing that I like working with the person that we pay from the open Collective. He is just such a great he's. Just so I don't know he just does such a great job, um so just getting to know more people and getting to work with people and talk about ideas. The work itself really feels like the reward to me. Yeah.

A

No I love that yeah I've I do feel like open.

A

S have been able to chat with before something like that like just meet great minds and you're like oh, my gosh, like even how we we met each other at the kcdc conference and I heard you talking about been Source. I was like wait a minute. What you're doing sounds really really cool, uh so yeah I I agree with that. um My other question I know that, like I'm kind of like maybe going off script, so I hope.

B

A

Cool my other question is: um um how do you think- and this might be a tough question so take your time to think about it. I've I do feel like open source and and Tech in general does lack a certain amount of diversity.

A

How do you think we can or what are ways that you're thinking or you think that people can increase that level of diversity and inclusion, especially in something like actual Actuarial science, because I already hear that like it sounds like it could be a struggle to onboard people, because you, you know you're busy right, so it can be a struggle to bring people in that.

A

Don't know much about this or know much about open source, so I'm, just thinking like just in general people from different backgrounds, whether it's race or gender or ability like what are your thoughts on.

B

That's that is a very hard question, and so there's there's kind of two two things that you know: um kind of I try to make them go together, but they're pretty separate and those two things are Actuarial science and open source yeah, and so you might have to approach that uh question from two different angles.

B

One is Actuarial science, and so that's really going to be driven by these professional organizations, which are large and well funded and have tens of thousands of members and they actively think about these things and they've been making changes to their credentialing processes, largely with a goal of trying to increase diversity, at least that's my understanding of the topic.

B

um You know. Basically, they they might want to uh I I, don't I, don't know a lot of these. The insurance industry is very centered in the midwest. As far as I'm aware and a lot of the big Actuarial universities would be uh University of Illinois, Champaign or University of Nebraska at Lincoln, and the surrounding areas are not particularly diverse and uh so I I. Don't know what to do about that.

B

Honestly, you have companies that are largely in the Midwest and you have universities that are in the Midwest and you have a bunch of you're going to have the demographics of the Midwest generally um and in terms of Open Source um I I, don't I, don't really I, don't really know. I mean the high school that I went to was very diverse and we had a lot of different types of people that were interested in science and I. I. Guess: I, guess that that question is out of my league that.

A

Is what I would say about that question? I asked I asked you a pretty intense question and I think that's a fair answer. I think, but I do think even things like this help with um improving diversity and inclusion like just being on the stream, because then you get other people who will get a chance to learn about what you're doing and your project like we already saw in the beginning, where you know people tuning in from Sweden and zombie and stuff like that.

A

So what you're doing already helps contribute I know: I threw you through.

B

You that's cool, that's cool, I'm, all full breath.

A

A

My open source questions to you. um Let me just check the chat if there's any extra like thoughts that people have here. Oh someone just said: um reinsurance models and underwriting related models. Okay I think they were talking about when you sh, you were demoing.

B

Well, yeah that I mean that sounds like some stuff that I haven't gotten into you.

A

B

So, that's why you know people that know things that I, don't know, really I, think to move things along. We might need a lot more involvement from senior Actuarial staff and they're going to need to actively try to do that transfer of Knowledge from the closed doors to the open world. So we we need insurance uh people that have been working in the industry for a long time to put a lot of effort into making everything more open. But I do my best with the knowledge that I have yeah.

A

You've done a lot you already, you already taught me a lot about Actuarial science I had no clue what exactly it was um someone's asking, I guess if you're, okay, how can people get connected with you, I.

B

Mean I'm on LinkedIn, um and so you can find me on LinkedIn. My name is as far as I'm aware, pretty unique, um so yeah Matthew, caceres and I'm on LinkedIn.

A

Yeah, let me and you're okay. If people connect with you, there I.

B

I don't mind if you send me a message that I don't know how to reply to then I may not reply, but if you send me a message that I you know so it just depends. Yeah.

A

I'll put these up and let me just remind people of different links. We got actuarials opensource.org, which is the website you developed and um I. Think there was a note somewhere of how how people can contribute I. Think you have written and I think you might have forgotten like you need playwright test and GitHub API requests.

B

Oh yeah yeah the basically our website sucks. That's that's the only way.

A

To sell it, our.

B

Website sucks it's built by a guy who has never professionally built websites before and built it on weekends and after work, so we in particular it's it's actually very hard. We have. We call the GitHub API a lot, because we fetch those stars and forks for certain UI components and um that those those API calls take time. So we need some sort of way of creating mock data, so that page loads are fast honestly, if you just try to make the website better.

B

Just if someone can make the website better and that's that that you need no nothing about Actuarial science to do that, just make the website stop sucking, make my code stop sucking and uh no I'm not asking for any features. I just want refactors quality of life, improvements and stuff, like that, and in my personal website, I, don't have a personal website. Actuarial open source.org is a website that I run I, guess I'll call it a personal website.

A

Yeah, no- and you did a great job, don't say- did a bad job. It.

B

Looks it looks good, but underneath is rotting and yeah.

A

Isn't that all cool faces I.

B

I I, don't know you know. I'm I haven't really worked as a professional developer, very much most of the lines of code that I've written have been on uh GitHub or on leak code. Yeah.

A

No I mean you do yeah, someone said like every other project, so, okay.

B

That's good to know well.

A

Someone said having interest in Actuarial science and machine learning along with Acca. This attracted my attention immediately. That's awesome, cool, um okay, so I feel like a lot of people said. This is a great session. I really did enjoy this as well.

A

um I loved, like just kind of seeing like how you can incorporate all these things, like bench, Lifetime, Insurance, I, didn't even know about all of this and you've like Incorporated it with GitHub action. I, think that was really educational to see um and just learning about your open source. Journey was amazing as well. I do want to just ask you some like fun questions that I normally ask for people.

A

um What's your favorite food.

B

Yeah, so my mom makes me oatmeal. She just made me oatmeal today, I live with my parents and uh she makes this oatmeal. It has almonds, it has pecans, it has blueberries and dates and apples, and it's just incredible and she makes a big thing of it, so that I don't have to worry about cooking stuff, and she puts a this Post-It note on it. That says so loved with a heart. Oh I know.

B

Most of the best I know yeah. So that's my favorite food.

A

I love it: okay, actually, okay, I added a couple other questions: what what but they're easy like? What's? The first programming language, you've ever learned.

B

um It would have been I, think C plus plus, which I didn't really learn, but they taught us in high school I joined a programming class and I had no interest in programming I. Just someone else kind of wrote me into it, and I wasn't very good at it, and in college I had a computer science minor, and there was some Java involved and again I didn't really like it and then once I could see that it solves real problems.

B

For me and I didn't have to use Microsoft Excel painful in Excel and in R it was just a joy that was when I said. Oh programming is a way for me to solve my own problems, and that was when a lot of stuff changed for me. So R was my first real programming experience. I would say yes.

A

A little bit when you, once you saw how applicable cool they're, just like so you're, going to make a linked list or like whatever and you're like why? Yes,.

A

Okay, wait another another issue: how will you like ideally spend your time like if you didn't need to work.

B

I would probably just try to I I think that without work, it's really hard to motivate myself to actually do open source work, because there's just nothing there's just like nothing really and so like once you have someone. That's like hey get this done, get this done, then you get into that mode and and it's like, but if, if I really was not so I guess, if I didn't need to work, I would still work.

B

If I wasn't able to actually work, then I would probably try to train for Endurance Sports at least 20 hours a week.

A

B

That point you'd probably be doing a like an Iron Man just because the diversity of training allows you to kind of avoid injury. So I I just been running 10K road races since January and I really enjoy it, but I'm a little injury prone. So I'm trying to work through some injuries and in the meantime, I've been swimming and riding a bike. A little bit and I still run, but I'd like to be at 30 to 40 miles per week and I can really only handle 15 right.

A

Now wow, that's amazing that you're trying to get to 30 to 40.

B

It's it's! uh It's aspirational, it's not impressive! It's it's aspirational, which is maybe like unimpressive to have people that aspire to things that don't achieve them, but that's how I am.

A

No I mean it's it's good to even just stay active because some people, you know once you start coding, you're like I, can't leave the computer, but it's good to stay active um okay. Last, like unfamiliar question, is like do you? How do you pronounce the word gif? Do you say Jeff for gif.

A

But it's not common. Most people are like GIF I, say GIF you're, like wow I'm, leaving no.

B

You just I bet you just say Jeff so that you can kind of mess with people you're like. Oh, if I say GIF, then it's like uh I, don't know: yeah like I, don't eat pizza just to tell people I, don't eat pizza, just to shock them, so I think it's the same thing with you and maybe saying Jeff I'm, not sure. No.

A

I, just I, just I've always said Jeff and then like I, started working in the tech industry and people like it's GIF I'm, like oh.

B

Okay; okay, all right all right, I'm, sorry for making for for questioning the authenticity of your pronunciation of gifts.

A

Someone said Jeff gang.

B

Okay, all right all right, um I got it.

A

um Wait there's like another I hope you don't mind answering this question. It's okay. If you don't know it, um someone said hi from the UK does actual open source.

B

Exactly yes, yes, that is the next thing that Cedric is working on so Cedric does this work and again it's four hours a week, um but you know we're working on trying to publish these benchmarks to the Journal of Open, Source software, or something like that and then after that, then he's- and you know when, when I'm done with school or whatever time, I can find uh we're very interested in simulated Life. Insurance data sets because there aren't any right now and I I just think it's.

B

It's really useful for reproducible research, because that data is proprietary, and so we need to do more research on methods and to do that. Research on methods, I think that we're going to want those data simulators. So that is a great question, and that is something that that I really think is important.

A

Yay I'm so glad that that was a good question. Okay, last but not least, question what is your favorite Beyonce song.

B

Yeah and so I, don't really listen to Beyonce albums I, listen to a lot of indie rock from the 2000s typically, but on 103.3 Hot, 103 jams in Kansas City. They have Beyonce songs, yeah and I, really liked that song cup it and then one day, I heard the song cuff it, but it sounded different and so I found out that it was actually the cover. The sorry, the cuff it wetter remix, and that song makes me feel like a baddie I.

A

Can't lie yeah, yeah I went to her her concert on Tuesday I felt like a daddy the whole time, I love it the great great song, Choice, um and also a great session I really enjoyed talking to you again. I want to say thanks to the people in the comments for always staying engaged and tuning in always brings me to like be like I gotta do open source Friday again, people are interested in it.

A

um So really love that, and then also thank you to you, Matthew for coming on sharing your time and your knowledge and and putting in this effort to help build out the the open source Community for actuaries and along with all the rest of your community. Who is helping to to build this out as well um for folks who are interested and want to check it out or want to help improve the website, go to actuaro, Actuarial opensworks.org, and then this is. This is Matthew's um LinkedIn, um but yeah. Don't don't don't spam him.

B

Yeah I, guess yeah and thank you uh for like having me on this is like such a cool opportunity and even preparing for it. I got the opportunity to just think about. You know the just to spend time thinking about you know what this is and stuff like that, so it's helped with a lot of perspective. Yeah awesome.

A

I'm glad I'm good all right, y'all I'm gonna um head off now, but bye.

B

B