Amir Bormand: [00:00:00] On this episode of the podcast I have with me Veno Raja. He is the Director of Engineering at Sure Prep. We’re gonna be talking about Azure Migration. Best practices gonna cover a variety of areas, uh, technical challenges, some of the upskilling challenges, um, finding the right people. How do you get the right people into help, and how do you know that?
Uh, ode, thank you, uh, for coming on the podcast and. Thank you very
Vinod Ralhan: much for inviting me. I really appreciate it. Uh, yes, all those topics are of my interest and I have been working in, uh, this industry for quite some time and last few years I have worked on few projects. Uh, so I would love to share my experiences and learnings.
Absolutely.
Amir Bormand: All right, before we get going, let’s, uh, start with two things. Um, one, uh, what does Sure prep do? So if everyone knows that, and then as the director of engineering, uh, what are some of your responsibilities? So Sure
Vinod Ralhan: prep, [00:01:00] we are in tax backend automation softwares. Uh, so if, let’s say you are filing your taxes, we have like, you know, only 15 paid binder.
It’s not a big deal. When the CCP funds filed for big clients, they have like, you know, 8,000 paid binders and that can be really big deal. Uh, so we try to automate as much as in the backend systems, uh, taking. OCRs to scan the system, verifying it, make sure all the numbers are correct, and matching it and trying to make all those automation, uh, so that it becomes easy and like gathering the documents, sending it to IRS, and integrating with the major, uh, tech softwares.
So all of those, uh, are being done and we have multiple patents. Use machine learning to make sure that, you know, everything is, even those things which we have not set up the system for machines can learn and they can, uh, uh, recognize those
Amir Bormand: documents. Awesome. Sounds like a pretty, uh, interesting product.
Um, and, and what’s in terms [00:02:00] of your, uh, director of engineering, what falls under your, uh, lap?
Vinod Ralhan: Yeah, sure. So I’m, uh, leading two products called SB and, and V scan. Uh, in that there is review results, so that those are the products I’m leading. And, uh, in those basically how you manage your, uh, Documents in the binder and how to arrange it.
Those things as well as how to verify, uh, throughout the process, uh, those are the things which are in my, uh, preview and I’m trying to migrate to the new, uh, new technologies, uh, in terms of taking to the Azure Cloud migration Also,
Amir Bormand: All right. Well, that’s a, that’s a good segue. Let’s, let’s dive into it. Um, so obviously we’re talking about, you know, you mentioned, um, uh, a migration, uh, to Azure Cloud.
So, so tell us how things look before the migration. Give us an idea of what, what the environment looked like.
Vinod Ralhan: Sure I will, uh, cover my previous company, uh, Meridian Link where I was doing it, where had, uh, [00:03:00] gone through the full cycle. So before the migration, everything was OnPrem. We were having our own servers.
And uh, uh, if there’s a problem we are dealing with that if there’s a, a spike on the load or we have to deal with that. Uh, so those were the challenges we were dealing. And once we did the migration, so we went from pure VMs, uh, basically pure hard, uh, servers. Then we, uh, went to, uh, container based systems and, uh, put those containers behind Kubernetes and that auto scaling and uh, uh, putting that, uh, as well as all the new technologies, uh, put in the firewalls and all those OS related issues, we were able.
And that way we made our life easier. Uh, so scaling was not a problem anymore. And uh, uh, maintenance was easier, although there were some new learnings and some new, uh, challenges [00:04:00] in terms of, uh, you know, how to navigate through the Azure portal and all those things. And how to, you know, make sure our logging and monitorings are.
Infrastructure is properly set up. Those are new things, but uh, overall I believe that, you know, we improved a lot and in, uh, in terms of reliability, uh, it increases a lot.
Amir Bormand: All right. Interesting. I, I guess you mentioned, um, You know, the, this is successful migration. In this case, we’re talking about your previous company.
When you look back and, and you look at some of maybe the, you know, the help that you needed, was this all done internal? Did you have to engage with the partner? Like, you know, how, how did you guys approach that?
Vinod Ralhan: So, uh, I will give first credit to, uh, our studio there. Basically, he, uh, took the initiative and he had, uh, good, uh, knowledge and uh, and we took the help.
Uh, we took the help from, uh, some, uh, consultants also. [00:05:00] And then we realized that, you know, Yes, we want to. So it, it has to be mixture. Uh, it, it cannot be one solution fits all. And like, you know, we went through multiple things. We went through training. We are told people to, uh, you know, learn. Uh, we provided the training model.
We have that, uh, LinkedIn, uh, learnings, uh, where everybody can learn from that, uh, as well as, uh, We, uh, got the consultant’s help and our IT help. They also, uh, did some certification and courses and they were very helpful into doing that. So this was like, you know, and we hired some already who were already expert in their area, hiring them.
So, Training your own resources, getting the hiring the best resources. Anyone taking the help of consultancies all come together. Uh, and that’s what made us successful. If you are gone to just one route, a consultant might help you in the day, and then later on you have to maintain it, then it’ll not have worked.
Uh, so [00:06:00] doing all three things, uh, is the solution I.
Amir Bormand: So I guess the, you know, you mentioned, you know, three, three different methods and obviously you wanna make sure that there isn’t a gap between, you know, the consulting team leaving and your team. How early in the process did you guys start planning the training for the internal staff?
So, obviously you don’t wanna be in a position where all of a sudden the consultants leave and you’re left with, you know, a solution that you may not know how to manage.
Vinod Ralhan: So very early in the cycle. So I would say even I, before even getting the consultant, so I joined that company in 2019, so I was.
Already working in the cloud, uh, in aws, uh, cloud, uh, before in the previous company. So my hiring was part of that. You know, my c also was new. Uh, all those hiring was done with the mind that we want to go to the cloud and then we, like CTO took the initiative and other people also, uh, took the initiative to make sure that we, [00:07:00] uh, uh, have the right people.
And we then start, uh, uh, training the other people also in the company. Uh, and we took, did all the resources and then we start hiring consultant. And tho that would then, would, would have then those were the very constructive, uh, sessions. If you, your own people are not, uh, not at a level where they, if they’re at zero level, those session will be like totally monologue.
The right question will not even come up. So you need to have your own people to have some level of expertise where they can actually ask the right questions.
Amir Bormand: And I guess in that sense, I mean, you have multiple parts that, that are moving to the cloud. There’s, there’s multiple aspects, security, development, you know, data infrastructure, um, yeah.
CL cloud, infrastructure A as you’re kind of looking at the various pieces that had to go to the cloud and, and looking at your team. Did you guys do this in, you know, was it all done in one [00:08:00] shot, meaning sequentially boom, boom, boom? Was it done in phases to make sure that, you know, everyone had a chance to kind of ramp up and, and make sure that one piece was implemented before the other?
I mean, how, how was the timing of, of the implementation of the migration?
Vinod Ralhan: So, no, it was not boom, boom, boom. It was, it is, uh, more than a year long process, I’d imagine. And, and, and some application were big and they, uh, would take time. Uh, and they were still in the process. By the time I left, uh, my application was small, so we became the first candidate in terms of who can go first.
And, uh, although being small, had advantage, but it still require a lot of work. Uh, so. There were different strateg like, you know, do you want to first like, you know, evaluate whether you wanna go to lift and shift, or we want to, uh, re-architect the whole thing, or we want to take intermittent solution. Uh, so we went to a little bit intermittent solution.
We said, okay. Uh, I had two application, my, uh, preview, which, uh, uh, migrated. Uh, one was, uh, we had been bringing donut core, so that [00:09:00] was good. So we pushed it to Linux containers and the other. Written way back and it was written VPN that, that we cannot migrate right away, uh, to Lenux container. And then we went to Windows Container in those cases.
And for those things, there was additional work needed. And then other things like, you know, uh, uh, you to, if you are going to the cloud, there are there things issue cams like, you know, You can get the containers in any time zone. So you cannot rely on the time on the servers on your prem. You can rely on, so you need to change your code a little bit to handle the dates correctly.
Those minor thin details need to be taken care. So there was some coding was needed, some was needed. Uh, how we want to as a firewall, right? We had it our, in our on prem, so we could. Which, which ips to allow, which not to allow. And now we are relying on, uh, web and then, uh, how, which we want to allow and not all [00:10:00] those learning need to be done before you make the migration.
So it, it is a lot of effort and, uh, it, it is. And then bigger application, bigger the changes and, uh, those need to be done and at the same time, You are still in the industry and you still need to develop the new features and you’re still competing with the other products and the competitors, so you cannot say, okay, I’m just doing the cloud migration and I’ll stop development.
That that’s not the solution. You still need to, to build
Amir Bormand: in parallel. Did the, I guess, you know, that’s a good point. You know, I’m making new features. As you’re planning for new features and you’re, and you’re going through the migration, obviously it’s, it’s not, you know, boom, boom, I’m done. It’s a, it’s a long process, you know, different phases.
How did the planning, uh, was planning for the roadmap affected? Did you guys have to make adjustments to, you know, accounting for this, uh,
Vinod Ralhan: It was effect to the most, but, uh, I believe, you know, we were a little bit more optimistic, so we, we were a [00:11:00] little bit behind. But I, I would, uh, take it as a part of growing pains.
You know, when you are doing something new, uh, you cannot plan everything as well. But we were happy with the result. That’s the main thing. Like, you know, although we did not meet that timeline, but we were happy with the results and, uh, the issues and scalability and those things, uh, we were able to make it.
So those were the good points
Amir Bormand: when you guys were, I, I guess the, the one thing that sometimes come to my mind when you’re. When you’re implementing something different than what you’ve had experienced, you guys made a hires, you know, for the cloud. You know, in preparation when you’re going through that interview process to, you know, interview people to make sure that’re the right fit, you know, I guess how did that, how did that interview process work?
Obviously you guys were, you know, not, maybe, maybe you guys had the internal, you know, team to already do the assessing of. Of those skillsets you needed, but how, how [00:12:00] did you guys assess people when you were hiring them to make sure that they were the right fit since you guys were yourselves still, you know, on a legacy.
Vinod Ralhan: So there was some, uh, good, uh, smart people in our company and like in those kind of interviews where we knew that these were the strategy position. So in one of the round of the, they will always be involved. So that was the part that we wanted to make sure that we hire smart people who we know that, who has some experience in cloud experience.
So, like my experience, I had. Aws, as I mentioned before that, and Azure was totally new when I came to Marine Link, and uh, that was new learning for me also, but having the cloud experience helps, uh, uh, jump into that, uh, train, uh, easily. And, and that’s the same thing. I try to make sure that we don’t need to have the exact experience which we want, but I look at the person.
Have the aptitude and some experience on those areas where they can at least, uh, uh, uh, have, [00:13:00] uh, the small learning curve and pick up quickly. That would be, uh, that thing. And then we start training our own in, uh, in house, uh, people, uh, so that they are not left behind too much, that they cannot even support it.
Amir Bormand: Absolutely. Yeah, I guess that that’s, that’s the flip side is, uh, you know, like we’ve been talking about the people that support it, making sure they have the right skills, and obviously the consultant team is gonna roll off and, and you wanna be, uh, self-sufficient. I, I guess if you’re looking back at that, uh, from the point in which the consulting team left and you guys were feeling comfortable, you could operate, you know, at maximum efficiency.
What, what, what kind of time did that?
Vinod Ralhan: Uh, so, uh, consultant left, I would say around, uh, uh, six to nine months kind of thing. Then like, you know, they were helping us and then, uh, uh, we were still, uh, a little bit, uh, you know, juggling through, but, uh, we were able to, uh, make it later on. [00:14:00]
Amir Bormand: It’s, uh, uh, I, I think these migrations always is about when, uh, you know, when people leave.
It’s, and, and you’ve gotta run things yourself and you have to be self-sufficient and, um, yeah, so you can pick up a phone and call somebody and, and get a question answered, but it’s different when, when it’s, when it’s your, um, area. I guess as you kind of look back, um, At that migration, does anything come to mind that you would’ve done differently?
I mean, looking back, I guess it’s kind of hard in real time to look at it, but you know, any lessons learned, anything you would’ve done differently at all? I think,
Vinod Ralhan: um, maybe, uh, we could have put a little bit more effort into, uh, training, uh, people, uh, earlier in the cycle, uh, and getting more so that everybody start jumping in.
Because in the beginning, only few people were, uh, working, uh, in the cloud migration and everybody was still into the features. Maybe we could have put more, uh, aggregate effort, uh, and. Do that way, [00:15:00] maybe that would’ve been a little bit, uh, uh, we would’ve get a little bit better results. But, uh, I see these things you can never predict.
Uh, you know, you learn as you go basically.
Amir Bormand: Absolutely. No, it’s, it’s, it seems it’s, it’s not a, it’s not a science. It’s more hard. I mean, you can’t possibly forecast all the, all the things, um, when, when you’re talking about training your team, and obviously they, they did. They did coursework and uh, they’re trying to, you know, do the lunch and learns.
Was there any opportunity for them to participate in the development when you were, you know, figuring out with the consulting team, you know, the strategy, um, was there opportunities to have your team hands on or was it a case where you really had to let the consulting, you know, take hold and manage engagement and, and kind of be a little hands off at that?
Vinod Ralhan: No, that, uh, our development team was very much hands-on involved. So, so our enterprise team was very good into guiding, and our IT team was [00:16:00] very good in that, uh, they were able to guide, uh, our, uh, development team. And, uh, we were, we were very
Amir Bormand: HandsOn. Did that, did that end up helping, like, you know, you mentioned the training beforehand.
Did that actually help accelerate their training? I mean, I know you mentioned, you know, obviously they were, Training beforehand. Obviously, you know, a lot of academics, but once you’re in the midst of it, did that help actually accelerate some of the learnings?
Vinod Ralhan: So what I would say, like, you know, the, the point where we missed, like not everybody was.
Like self-motivate and, uh, curious to, you know, uh, guide themselves. So then like everybody jumped, uh, on the boat later on. But in the beginning some people were still, uh, doing their own, uh, function development, but not on the Azure, uh, development. So that, that was like, you know, there’s personal curiosity and some people are curious and some people are not.
So, uh, the point was like, Then like we had, we were dealing with a little bit less bandwidth who were, [00:17:00] uh, more curious and uh, uh, in the development. Um, so that’s what I was saying, that, you know, in the end everybody was on board, but in the beginning it, that that transition was little slower and that, that took us a little bit extra time.
Absolutely.
Amir Bormand: Well, I mean, it’s, it’s, it’s, they’re, they’re not easy. Uh, when whenever you’re migrating anything and, and you talk about new skillset sets and you know, the timelines, you know, the work still has to be produced. I mean, it’s a very convoluted, uh, project. Sounds like it, it, it worked out. It sounds like, I mean, you know, you, you did, you know, did the best plan you could.
Uh, I appreciate you coming on and sharing cuz um, these migrations are not easy and, and I think, uh, bringing visibility to, to some of the thought processes always is, Thank you. Absolutely. Um, before I let you go, I was going to ask you, uh, I’d like to ask two final questions before you go. One is, you know, if a future guest, um, you could ask them to do a topic or cover a specific [00:18:00] question that you might have.
Is there anything that you’d like to hear about?
Vinod Ralhan: Yes, I would like to, you know, uh, recommend AJ for, uh, he’s the director of engineering, senior director of engineering in Visa. Uh, I can actually check with him if he’s available. And, uh, I, he would’ve some, uh, very good points because they are running the whole cloud in-house because, uh, visa do not, uh, subscribe to the public cloud.
So they’re, they are doing the whole cloud things in-house. So that would be interesting. I would love to hear. Okay.
Amir Bormand: Uh, right. Yeah. If, if you know somebody who’s, uh, uh, running the cl you know, running their own cloud, uh, in-house, and, and, and he’s, he’s a willing to be a guest, to be on a talk about that. I think a lot of people like to hear about that.
So I, that’d be fantastic. Um, if somebody wants to reach out to you to, you know, touch base with you on anything you mentioned on the podcast, but what is a good way of getting ahold of you?
Vinod Ralhan: Uh, you can reach through my, uh, Gmail, gmail dot. Or, or send me a [00:19:00] message on connecting with me on LinkedIn.
Amir Bormand: All right.
We’ll make sure we include, uh, those links on, on, uh, the show notes. Uh, Buto, thanks for coming on. Thanks for sharing. I appreciate it. Thank you. Absolutely. That’s it for this topic. We’ll back, yeah. Different topic, different guests until, and, uh, two things. Uh, in this case, uh, we’re gonna see, hopefully VE node can help us, uh, uh, coordinate, uh, having somebody talk about in-house, uh, cloud, uh, infrastructure and how that’s, uh, being managed.
I think it’s gonna be exciting topic, um, if the podcast is useful. If you enjoyed it, please share it with somebody. Um, if you can leave a rating, uh, on the Apple Podcast or whatever your preferred platform is, uh, that you listen to the podcast on, that’s how we grow. So I can’t thank you enough for doing that.
Otherwise, that’s it for now. Be back again. Thank you and goodbye.