Revolutionizing Housing with AI

Full transcript

Ai's industry impact

Welcome back to the podcast, guys. Today, we are joined by Josh Blair. Josh, welcome to the podcast. Thanks for having me. So why don't we start this by you telling us a little bit about yourself and what you guys do at Kensho Software? Sure. So I am a lead software engineer for Kensho Software. We're located in Denver, Colorado in The US, and we build a SaaS product that is used by housing authorities, both tribal, Native American tribal housing authorities, and also public housing authorities. So I I kind of ran some numbers just to have some some data for you, and I think that we are serving about over a 100,000 units, so we have we have over 200 housing authorities as customers, and across all those housing authorities, think they're roughly, I don't know, between a 100 and a 150,000 units that are under our management, and our software helps housing authorities manage all their data as far as programs, projects, units, households for tenants, the household makeup, it handles rent calculations, mortgage calculations, it handles all the payments and transactions for households, as well as all the

vendors for the housing authority, so the housing authority hires vendors, so we have that that angle of the business. Our software helps with managing their inspections, so unit inspections, and also when when there's an issue with the unit, typically a work order will be generated from that, and we have we have a SaaS product that is web based, We also have mobile apps that help with inspections and work orders. Well, that's a lot of use cases, Josh, I must say. What's the biggest trends that you're seeing right now across software industry? Well, as far as software in general, a lot of shakeup in the in the industry. I see a lot of, you know, tons and tons of stuff about AI, AI's huge, it's changing by the minute, each company is leapfrogging each other, you know, trying to trying to gain that foothold. As soon as something great comes out, something else beats it, it's it's it's it's in high transition. We're we're just trying to make sense of it, we're trying to leverage it to our advantage and to our customers advantages. And then with with AI and the current economy, it seems

the tech market is is a tough spot to be in for people looking for jobs, people losing their jobs through mass layoffs. So that has got me definitely a bit concerned. Mhmm. And like my following question to that would be, how are you going to leverage AI in, you know, your software that you guys are developing? So we we do have a couple initiatives currently kind of ramping up. We we did build a an AI a genetic AI tool in our our in some internal software that we have, it was a proof of concept, and it worked really well, we used Rag to consume lots of console specific documents, as well as also lots of HUD, so by the way our software helps housing authorities be transparent to, and adhere to the requirements of HUD. HUD stands for Housing and Urban Development. So we were able to add a lot of the HUD content to our RAC, and we could very quickly get answers that would have taken a lot of time for an individual to sift through all

AI as a team member

all this this government documentation, so that was that was a nice benefit. So I think that we could do more projects like that for our customers, where where we can provide instant access to data that would be hard to to to gather just just with normal search, normal normal tools that we're all used to, I think it's it's gonna become much more mainstream for a lot of products to have this this kind of stuff built in. And then from a dev team, or our tech team I should say, our developers, our data ops team, and our QA are all leveraging AgenTeq AI heavily. Mhmm. Is there any interesting technical challenge that you faced quite recently and how you tackled tackled it? I I I run into that every day. This morning was was no exception. I did a a late night deployment last night for a new feature, very highly requested feature, I got everything verified and dev, I pushed it to QA, our QA went through it a number of times, there was a little bit of back and forth, and then once QA signed off on it, we would accept its process.

It made it all through all the way through all of our processes, and it got deployed, and there was an issue. And it was it was an issue that I I I'm using AI to help you with right now, like right before we we joined this call, AI is helping me determine what went wrong. So, you know, from that standpoint, it it it I use it as a pair programmer sometimes, I use it as, you know, a jigsaw to build stuff, but also to to double check me, you know, to to kind of be a a checks and balances on what I'm doing. So I find that very helpful. And like like some tech companies, not all of our software has coverage, and I want to use AgenTeq AI to help round that out, you know, completely, so that we can we can do big deployments like this, and have much more confidence.

I mean, and this is a pretty rare exception, but it it was it was very timely that you asked that question, like, how does AI help you? Well, it's like it's like on the team members sometimes, you know, it can can look at something I've done and say, well, you know, I I think the issue's around here, and then I can I can investigate further? So it's it's been super helpful for me, I use it a ton, my my team is is getting so good at it, we're definitely seeing some gains from both quality and output. It doesn't always save a ton of time on a on a large project, but it definitely helps me build I think more more more comprehensive solutions, it it'll it'll it'll help me to surface edge cases that I might not have thought of, you know, so it it can actually save some of the some of the back and forth, so that that feedback loop that that could be expensive from a time standpoint and and resource standpoint, I think it's it's it's cutting that down a lot. So, yeah, pretty

Finding the right tool

pretty pretty excited about it. What what technological stack are you running in terms of, like, AI? Do you use quote code? Or Great question. So Codecs. We we we currently have a team team license set up for Codecs. I've been using Windsurf, which is a different kind of a tool, but it works in both the editors that I use, I use both Visual Studio based editor and JetBrains Writer, because all all my back end is in C, and we're an AWS shop, an Amazon Web Services shop, so I build a lot of Lambda API RESTful endpoints and GraphQL endpoints, and C, and so the Windsurf plugin works in Ryder, and Windsurf forked a version of Visual Studio, sorry, Versus Code, and that has Rider built into it, so I use both of those exclusively. I am I am playing with Cloud Code, I'm I'm I'm trying I'm trying Codecs too, because because the team kind of settled on that, but I was a bigger fan of some of the other ones, like Cloud Code and OneSurf.

I also do so much AWS work, I I'm also working with Cairo, I don't know if you've heard of Cairo yet, but it's another one that's it's it's put out by AWS. Mhmm. I always like to ask you this question. Josh, if you had a magic wand, right, what would you change about your guys' industry or just cancel in general? That's a great question. Let me let me think about that a little bit. We have a very small team, and I would I would love if we each had a little more time for each project, so that we could spend time thinking about the Optone solution, you know, instead of being so constrained with tight deadlines, and very tight resources. We our team is a total of six developers. Mhmm. We have we have enough work for 40 developers, and we're trying to we're trying to cram eight pounds of stuff in a two pound sack, you know what I mean? So it would be great, I think for for the for everyone on the team if if we we could just take a little bit more time on each project,

that would help help us build better solutions, and also help to ramp up on some new concept. You know, instead of instead of rushing through it, getting it done, to be able to spend time and really really absorb the the the complex, the complexity that we're fighting with, that that would be nice I think. AI is helping, certainly. I think we have a few more big features that we're working on, and as soon as we get those to the finish line, that's supposed to help us unlock a lot of revenue, and I hope with that we would we would probably bring on some more resources. We do a lot of mentoring, and and and the team is very cohesive, so we we help each other a ton. As as the lead software developer, I I help all all the developers, but Junior has helped me all the time. I there's there's there's no there's no pecking order, we're we're just we're just trying to all be productive, be be helpful to one another.

I love that part, would love to be able to to to grow the team, maybe maybe double it, I would I would say, maybe maybe specialise in front end and back end, because right now we're all we're all stack. Yeah. Anyway, that's that's that's what I'm thinking. Mhmm. As a lead software engineer, could you say, like, what kind of skills do you think are the most essential for your role specifically? Sure. So we we have been in a modernization phase for over five years, maybe six or seven years. So number one, having having a understanding of the entire the entire stack, all the old stuff and the new stuff, we're breaking apart a monolith, so that's part of our modernisation strategy, so we're using kind of the strangler pattern, and and building new stuff in a in a modern way, we're we're transitioning older features into that that new paradigm, we're scaling out in AWS, which is is helping with our our larger customers, specifically some of the public housing authorities are quite large. So that is a that's a big effort too, and just be able to to keep your eye on on

the, you know, surface level, and then, you know, that that higher higher level view, having a comprehensive understanding of how everything works together, being able to troubleshoot, being able to look at a problem, and apply some of the architectural patterns that we have already put in place, the right way, you know, the right tool for the job basically, that's that's a that's a skill because, you know, a lot of things can be done multiple ways, but some ways will bite you weeks or months or years later. And, you know, we're we're learning from our own mistakes, honestly. That was a very, very good answer, I think, actually. Like Thank you. For people that are listening right now and would like to check you out or maybe cancel, actually, where should they go? Well, I can reach the Kanso website at at @kansosoftware.com. Kanso's spelled k a n s o. It's a Japanese word that means simple. So so our software, the whole idea behind our software is is to make this very complex problem domain, or business domain easier to deal with. So the people that use our software are typically serving those

those individuals who need assistance, know, they're they're helping with low income housing, and and vouchers, and you know, all these different programmes, and and basically the software that we build helps helps those people help those who need need a lot of help, so that that was a hard a terrible way to say it. We serve those who serve the underserved. How about that? Yeah. Lost my train of thought. Sorry. Oh, it's okay. Well, thank you, Josh, for coming on the podcast and spreading words about cancel, and we'll see you in the next one. Absolutely. It was great to meet you. Thanks so much.