How to Start With AI When You've Already Decided It's Too Hard (13 Year Google Veteran POV) - w/ Aishwarya transcript

But when I hit my 12th anniversary, I was thinking about, hey, I've done so many things at Google, but what does the world look like outside? Uh, so I left Google. Google. 13 years at Google, YouTube monetization. She walked away in the middle of the biggest hiring shift in tech. tech. I was using Gemini internally at work, but when I wanted to use Claude or something else, I was not using it at work. I was kind of tinkering on the outside. So I wanted to be in a place where I could pick the model I wanted and then you know I didn't get that job. There are layoffs happening and you know not getting this one job felt like wow this is going to be a very tough market. And then she landed. By the end of this episode I promise you'll know what companies actually mean when they say they're hiring for AI roles and what to build first.

Companies are hiring for AI skills but you have to read between the lines in terms of what they're hiring for. People think about learning AI, everyone thinks about the agentic use cases and think, "Oh man, that's so difficult. I'm not there yet." And don't even try the first few steps to get there. 13 years inside, four levels out. Please join me in welcoming Ashwaryia Ragavan to the Rei Setu podcast. Let's go. Ashwaria, welcome. Thanks, Amanda. Great to be here.

You spent 13 years at Google mostly in YouTube but also outside in towards the beginning of your career and then you chose to walk away from that and you are now a principal product manager at uh Aria. So do you mind kind of walking us through the why behind that kind of how that all unfolded for you? Yeah, for sure. And uh we might be touching on a little bit of that further too. But the summary of it is um I had joined Google 13 years ago when uh uh I got to do a lot of exciting things. Uh one of my uh first experiences was uh building machine learning inside Google and uh uh tackling some of our biggest problems. And then um the nice thing about Google was every time I was itching for something new, there was always an opportunity inside for me to pivot into. And I had done multiple pivots inside Google. But when I hit my

uh 12th anniversary, I was thinking about, hey, I've done so many things at Google, but what does the world look like outside? And that got me curious. And uh so I left Google one to again figure out uh what it was in the world outside that I could do. Uh but second and you know I think we'll touch upon it a little bit more there was a lot happening with AI um in the world and I wanted to have a free reign to explore it in whichever way I wanted and didn't want to uh be uh restricted with what I could do just internally at Google.

So that makes a lot of sense but I think where my mind is going next is it it almost feels like a paradox a little bit right. So what I mean by that is obviously as we know Google is at they have one of the frontier models. Gemini I've heard arguably is actually even better than claude and open AI when it comes to generalist AI or/nowledge work right. So it's not that there wasn't um interesting work at least in AI being done at Google. So for those of us that don't understand how big tech works that includes me do you mind sharing a little bit more about why you still felt restricted? So does that question make sense? sense? Yeah, it makes sense. Again, leaving Google was not at all an easy decision.

I spoke extensively to people internally at Google as well as outside to make this big leap of faith. But I think there are maybe two things to think about over here. So one was at Google again I worked in YouTube and um I worked on um monetization and so in term and and so from doing that to switching into AI at Google just meant I had to do uh go through a lot more steps um and again Google is such a competitive place there were so many people who are also trying to follow all those steps. So I think that was one of my realization that um you know again like the world is bigger than just Google and there are going to be probably more exciting opportunities if I didn't constrain myself to just the Google ecosystem was one and second to your point about um uh models itself you know you're right uh I think every few months our preference

for models changes like for a few months it's open AI and then it's Gemini and then in the last six months For me, it's been Claude and I think everyone's kind of going through that journey. For me at Google, I felt like um I was using Gemini internally at work, but when I wanted to use Claude or something else, I was not using it at work. I was kind of tinkering on the outside. So, I wanted to be in a place uh where I could pick the model I wanted and you know like make changes as things evolved and didn't have to wait for any specific like permission to do that. Yeah, that makes a lot of sense. Um, and now to take I guess a step back, you've spent so much time there. You started off in Hyderabad, ended up in San Francisco over the years. Can you like I know this is easier said than done, but I'm so

curious to kind of unpack not just your journey at Google for 10 plus years, but even more so from a lens of what a job in big tech looked like pre-AI era, at least at the time it was called machine learning, right? referred was the you know the cool AI term was slapped on it but um and again I can obviously fill in some of the major blanks where of course you're not writing code from scratch and such but yeah if you can kind of take us into how you saw the world evolve like so if you know just to take a pause in my case right so when I started this podcast 2 years ago basically every single thing I did all of my workflows are a 180° from what they are today given where AI has evolved just in the past two years. So I can't even, you know, wrap my mind around just the sheer amount of volume of change that you

would have seen. So I realize there's many different ways to attack this, but yeah, just kind of as an open-ended one, how would you kind of describe the transition at Google or your journey at Google over the, you know, your entire time there? Yeah, I think it might be more relevant to talk about my journey because again Google as a company has so many different products and each kind of product has had its own journey and how every product uses a AI or machine learning is kind of beyond me. Uh yeah but just sticking to my kind of experience at Google. Uh so I actually graduated during uh the 2009 recession and for uh the first two years after that I didn't actually have anything uh strong that I was working on and then I worked in this analytics company based out of Chennai for a couple of years that gave me the foundation for

analytics and machine learning and again this was the nency of data science analytics machine learning whatever we we call this um and and then when I joined in Google. Uh I had joined this division which was tackling a lot of um longtail problems that people didn't have the bandwidth to give attention to and this is the kind of problem that is very ripe for machine learning because again machines can make decisions at scale. So, uh, my colleague and I, we took this as a side project to see, hey, can we just throw a little bit of code at this problem and see if it makes decisions at the same quality as human but at better scale and we had some very early wins and very like uh strong wins too and and I think for me that completely changed the perspective of what you can do with whatever we call it AI today or machine learning but

doing things at scale. Um and another learning that it gave me was I hadn't studied uh machine learning or analytics or any of it formally in my college because uh analytics as a field only started evolving as something mainstream after 2 three years after I graduated. So for me it was always it always meant I was learning on the job and uh kind of keeping up with what was coming up and so um and so that was kind of my introduction to uh using machine learning at scale at Google. Um and then after that I joined the finance team uh again because the I the same problem set exists in finance too. there's a lot of data we're trying to make decisions and uh the scale way to make decisions is being smart about it. So the same theme kind of followed and I see the same thing again happening with AI now it's

very different way of thinking obviously uh because uh like the machines are can now learn a little bit more and they are smarter and you know to your point our workflows have completely changed but some of the basics still stays um when a machine gives you an output you still need to review and see is this reasonable or is this not which means you have to in your background know what does good look like what does bad look Right. And so I think some of that basics still remains. The tools have changed but I think the foundations remain. remain. Yeah, that's very interesting. And actually I had an Amazon senior data scientist on the show that that actually said something very very similar and they were like ultimately whether you're using cloud code or if you're just you know writing code by hand or you're you

know just using like a TensorFlow pre-built u model it's not called a model it's framework right. Yeah. So um they were like it doesn't matter as long as you understand what it is that you're trying to do. You understand what the input is output is and how you're getting from A to B you will probably end up being fine. And it's one of those themes that keeps coming up again and again where people are like you know people such as yourself um you know such high pedigree people which with that have such great years of experience at big tech orgs and I hear this theme again and again. So yeah, just wanted to call that out that um yeah, great minds think alike is I guess what I'm getting at. Just to kind of shift gears into um the hiring side of things because obviously you you just know big tech from the inside so well and so deeply.

Can you maybe help set the stage a little bit for folks that are currently trying to get into we'll just say tech, right? It doesn't even need to be big tech at this point. for those that are trying to get hired into um AI roles we can say right as of today today um what would be some of the you know most important um traits I want to say or skills that they have that are underrated right cuz I'm sure everybody realizes you need to you know be good at lead code you need to be good at your data data structures and stuff like that but in your experience are there any like super underrated things that have a lot of value when it comes to this.

Yeah, for sure. And this could apply to any role not specific to data, but I think the main uh the one of the biggest differentiators in interview is the storytelling of uh of the storytelling and communication aspect. Um let me maybe uh provide an example. that would be great. Yeah. So I I think everyone's familiar with the STAR framework and uh many of the well-known frameworks but I think what people end up doing in an interview again because it's a very nervous experience going through an interview is you just go through it as a checklist.

Uh but instead if you can kind of put yourself in the shoes of the listener or the hiring manager or your interviewer, they're asking you that question because they have a specific problem in mind. they are probably experiencing it in their work right now or they've had a bad experience if it's a behavioral question. it's a they've had a bad experience with uh a a previous colleague that they are kind of trying to say hey how would you deal with the situation so I think being able to tell the story in a way where you understand the challenge you actually talk about what you've done and then talking about the impact so thinking of it as a story arc I think is a very underrated skill because the more you can engage the interviewer instead of checking a box I think the more they are going to see you favorably as they think about the

interview experience. Um the second one um in similar lines is also uh really anchoring on the impact and so what does it mean for the interviewer? So what that you launch this product for your team, so what that you resolve this problem for someone. So being able to again like contextualize your uh your successes in your previous role with what they are right now hiring for I think would be one of those again overlooked things. Yeah, I like that because and it's so interesting because I actually have always heard of the so what framework when it pertains to résumés. So every time you're writing your resume bullets, every single of your work experience bullet must share the so what right?

Like we see bullets like drove or built such and such dashboard that you know did reported on such and such but so what right like what was the impact that it brought about so that I was familiar with but that's such a cool extension of that same idea where yeah the same thing does apply to an interview as well cuz yeah it's not always immediately obvious in terms of you know what what is being um communicated around the behavior vir behav vural uh uh tip that you gave. So a couple just couple additions to that that I have found that helped sometimes was first there's such a thing as apparently a star L, right? So the L stands Yeah. Yeah. So the L stands for learning learning which is which is very LinkedIn coded but it it it it does make sense when you think about it where it's like a nice little cherry on top where it's not

expected and when if when you leave that and it's usually the last thing you say after you've gone through your start I found that you know and from just not just my own experience but other kind of mentees that I just talked to and kind of help around this type of stuff they say that they they do get pretty good feedback around uh when they do implement that. So yeah, just kind of wanted to tack that on to what you had said. said. That totally makes a lot of sense and especially now, right? I think what my my sense of what people are hiring for is um there is kind of a blurring of roles happening right now. Uh there is a blurring of ownership and accountability. And so a lot of the behavioral questions the way I interpret them is saying hey we are hiring with these kind of um uh assumptions about your role and another team member's role

but as they evolve how are how resilient are you going to be and this is why if you have past experiences of dealing with changes or conflict or negotiating or dealing with trade-offs I think usually the more successful story say I did this incorrect correctly. I learned this and then I fixed this and this was the this this was the result. So having that whole arc of again it's like the hero's journey of there was the struggle, I failed, then I learned and then I succeeded I think really helps.

I want to double click on what you just said around the blurring of rules that's happening because a I think I kind of get it but for those of my listeners that are not super into like Silicon Valley trends which half includes me, can you share more about that? And I guess what do you mean by that and how is it happening right now? I might be uh referring to just the subset of people I come across uh uh I come across uh from um so it might be a small sample size of again what I'm sharing. sharing. Yeah, I can sorry to interrupt but I can share that it's not because I have actually heard this from other places and like podcasts and such as well. I I've just never had somebody help break it down for me. So yeah, I I don't think it's like a small thing. I I do think it's happening right now in real time if we can put it that way.

Yeah. Uh so maybe two concrete examples I can share from uh my experience is I mean like even like two three years ago a data scientist's role would end with making recommendations. you do all this analysis and you say okay this is my uh this is my um my recommendation but now again this is where the blurring of of roles the the consumers of the data scientists decisions can now go to claude and kind of do some of this lightweight analysis themselves and so they are going to again start making some decisions and again depending on the company there might be some tearing of if a decision is more than this magnitude then it has to go through a minted data scientist. But if it's maybe a low-risk decision, then you know you you should self-s serve. This might be kind of how some companies are positioned. So again, a product

manager's role becomes being able to do a little bit of what was previously data science's role to now that you know it is now also part of their job. Uh so that is one example of blurring. Mhm. Um, another example, and I I don't have a direct experience of this, so I'm quoting uh one of Lenny's podcasts guests um who was sharing something on the lines of um lot of again, especially in big tech and a lot of large companies. Many people's role exists in passing information from one set of stakeholders to another set of stakeholders because again, these are very complex team. They're very matrixed and uh it's it's hard for one person to hold all the context in their head about five different products and how it might affect a launch that they are doing. So again this is how the system work but now the hypothesis and I don't know if

this is happening in reality is you should have AI agents that are listening to all these meetings summarizing and bringing to you this context that was previously managed by five people um and and so again that's where the blurring of role happens like so what used to take five people to do should be done by one person and maybe in half the time for them. So this is maybe another like u example of blurring of roles.

That's very fascinating. Yeah, I was definitely not aware of the AI agents one mostly because I don't have the faintest clue in terms of how big tech operates and the type of roles that they have. But I am curious given your background and your just you know your career trajectory where do you personally stand in terms of that change when it pertains to the AI agents? Do you think that we are there already where theoretically this would be a good enough idea for this to just work reliably or do you think that that's still a ways down the road or do you think that that's actually never ever a good idea? Uh so I have a small newsletter and uh I had a an ex-colague of mine who was at uh Google and then Meta who implemented some of these AI agents for her day-to-day tasks. So again her role was a technical program

manager which meant she had to coordinate between multiple engineering teams and make sure if one team is blocking another team provide an update and also managing uh a weekly executive uh reporting or executive readout on hey here is the status of the things we were developing this week. Here are things that proceeded. So uh essentially managing this big product machinery and uh she shared that she had uh built these AI agents that kind of again um are part of all these meetings have updates that they're automatically like feeding to her uh and she has an open claw like product that pings everyone automatically and says hey you haven't provided an update or we haven't seen any movement on this tracker is this still on track and so I I again um I haven't implemented this myself and I'm super inspired to do it for my own job,

but hearing her experience shows me that I think people are trying it out and uh she was sharing that it cut down 80% of the time she was doing these tasks previously. previously. That's so interesting. I I actually can't help but feel like the the thing that I guess the what might derail this for me is if I have a bot texting me, I'm just going to ignore it. Does that make sense at all? Do do you feel that or does it not work like that?

I I know what you mean. Uh I think the again and this is where you know the whole it's not just having that bot set up. It's the systems around it. I think in her case if the bot didn't give her an update, she's going to say this project is at risk. And if that goes to your VP that this project is at risk, then the person responsible for giving you that response is going to respond to the bot. So uh I think this is where it's and and I'm and again I I don't want to make it sound like she implemented this lightly. I'm sure she set a lot of guardrails around this and made sure that everyone understands the implications and the incentives around why we are doing this or why she was doing it. So uh it's it's a lot of change management but I think it's also one of those things where you know as a person I don't want to again like bug 20

people to get updates if I can have someone else do it for me. Uh yeah, next thing we know, we'll have a bunch of people's agents all talking with each other and then Yeah, right. Cuz I mean that feels like a good idea to me, right? Why do we not already have that? Just remove that waiting for somebody's ping or waiting for somebody's email. If they have the information, just have their agent relay it to us. Like why why are we doing this song and dance that serves no purpose?

Exactly. And even for you and me, right, when it took us so many exchanges to make this happen. If my agent spoke to your agent, this should have, you know, at least saved us all those like back and forth messages. So, no. Yeah, that's a good point for somebody that's trying to um just upskill more upskill themselves more than more in AI. And for context, let me yeah, that was not a good way to jump into that at all. But I have a friend, he has a degree in data science.

He's currently looking for jobs for whatever reason outside of you know trying to teach himself new models or implement projects and such. He just has not found found the need to get super savvy when it comes to I mean at this point at least at the time of recording this in early June 26th. Maybe you can correct me if I'm wrong, but there's really no big difference in my head between like an open claw or Hermes or Clot Code or Codeex. Like they're really all the same. Dare I say that and but with the floor open for you to critique that. But what I'm getting to is that yeah, he just kind of minds his own business except he feels that it's getting in the way of him landing a job.

So that I know this apologies for bundling like three questions into one, but first that and then second, do you think that being an agent orchestrator is kind of a must-have skill for most tech jobs right now in 2026 or can you still get by without being super into the weeds of this stuff? Uh so let me tackle one question at a time. So let's say is the first question um um are they the same? Are are these things basically the same things or do you are are they different in your head?

Yeah. Uh I and this is you know kind of what we were alluding to earlier which is these models keep changing like even my own preference of last year uh I in fact have um like GitHub and other things written with Gemini because that was the platform I was most familiar with. with. Same. And it was the best at the time cuz literally this time last year there was nothing like Gemini 3 Pro when it had first dropped. I remember how insane it was. I used it until like December and I haven't used it since cuz it just you know for me at least just fell off a cliff. cliff. For me I think the transition away from Gemini to Claude code came with all the hardness that Claude started having with you know you could have memory you could do everything in the terminal. it also start became very easy to talk to multiple tools etc. So and now I think

it's becoming the industry standard um every AI lab and uh I think even Microsoft recently and basically all the companies are now figuring out that you can't just sell AI uh power users want all these integrations and then that's going to become table stakes and the next thing is going to come up and even open claw wasn't a thing until December right or Yep. So that's right. Yep. So I think uh my point of view is we can't anchor so much on one specific tool and uh become fanboys or fan girls of one tool and these things are going to keep evolving and sometimes something that worked for you again you know to your Gemini example last year you might completely forget and move on. So I think that's just going to be how it is.

Um is your second question does everyone need to be an agent orchestrator? Yes exactly. Yep. Nailed it. Uh I wouldn't say everyone because uh that's I think um even like my my experience at Google six months ago it was not an expectation to do anything in AI. I think if you did something in AI cool you know you're saving some time for yourself but I know right now we are seeing companies like Amazon have a token leaderboard and I think I think it's it's evolving. Um so I wouldn't say agent orchestration is a must but I would say that um companies are um again this could be my small sample size.

Companies are hiring for AI skills but you have to read between the lines in terms of what they're hiring for. Some companies when they say AI skills I think they mean basic things like are you using Chad GPT? Are you using some form of beyond just using chhat GPT as a search engine? Are you doing slightly more automated tasks in charge GPT like do you have a repeatable process in charge GPT? And again I wouldn't say this is bulk of the hiring. I think there are some roles that are saying this uh like for from roles I was previously like uh looking at or I came across were roles in finance or HR or some of these traditionally non-technical departments like some proficiency around um around AI usage was a trend um I think the agent orchestration is maybe more on the far end of what is kind of mandatory uh I

think having that definitely will differentiate you and if you have that skill, I think it's a really really rare skill. So, you make sure to find the right company and get those roles. Um, but I and I say that with a caution because there are times where teams advertise needing this agent agentic skill, but then you have to interview to understand do do they understand what they are hiring for? Is this a checkbox for them or do they um or do they have the culture and the processes in place to adopt AI? Um and the reason I'm sharing this is um if if the company as a whole isn't fully bought into again these agentic process etc but you're hired to do that then a big part of your role is also going to be that change management and uh um all the human aspects of you know negotiating and convincing and advocating and for some people that

might be a great role but knowing that your skill is not only going to be the technical AI skills but also all these other human skills is something that you have to know when you get into those roles. roles. Yeah, I think that's a great call out. I am surprised when you said that it's a rarer skill to be an agent orchestrator, I guess. So, just so just to make sure we're speaking the same language when you say agent orchestrator, maybe what are one or two tasks that you are expected to, you know, like just one shot or just know how to do. So just so just to make sure that I'm understanding this right because that is a little bit surprising to me because isn't everybody an agent orchestrator now?

Yeah. Yeah. Uh and this is where I'm I I was speaking from my sample size and uh the the example that we were talking about uh like my ex-colague who has this agent that pings her teammates when they don't update their status etc. Uh I I've seen I mean I've I've spoken to other product managers and tech etc. I don't think everyone's implementing agents. I think some of them are maybe in more uh advanced stages of experimenting or implementing and some people are in you know are continuing to do their traditional PM job with again they're using AI for some of the more repetitive tasks again like AI for data analysis instead of uh having to work with a data scientist etc. So, and and that's why I don't want to say that you absolutely must be an agent orchestrator to get these jobs because there are multiple

flavors of this. And uh the agent orchestrator I had in mind was the example I shared. I I think it's definitely a useful skill to have. I wouldn't sleep on it, but at least again in June 2026, I don't think every single role mandates it, but having it will definitely put you in a higher um higher tier in terms of candidates than not having it for sure. Um what according to you would be the best way for somebody that's been putting away learning these things, they've realized that hey, I need to do this like yesterday. They haven't. It's been 6 months. They're listening to us right now. What would your advice be to them in terms of trying to get them to I mean they don't have to be an expert like you said but at least to get to a point where they can have an educated you know back and forth conversation

with somebody that understands this stuff. stuff. Yeah. Um, I mean, I I'm sure everyone's been hearing this, but actually, you know, doing things hands-on is going to be the the only way to learn this because there's only so much theory you can get from what you read online or watching YouTube videos. So, definitely trying something hands-on and and this is where you know there are I I think of there are four levels to try AI. The first is using Chad GPT as is. And even with chajp, you know, the more like if I ask chip to write an email for me, it's going to use words that I don't typically use or it's going to be overly formal or overly polite and not have the context. So just figuring out baseline, how do I have chat GPT actually sound like me or you know any similar use case. I think that's the first level.

The second level is making this more automated. So I again I'm taking the example of CH GPT because that's usually everyone's consumer consumer product. Uh but in Chad GPT itself uh I forget it what it's called. There's a way to um kind of have a custom chat GPT. I should know the name for this. Yeah. Do you mean a custom GPT? That's what Yeah. Yeah. Yeah. Custom GPT to to do repetitive tasks like you know it could be writing email, it could be polishing your resume. So having that custom GPT and again that repeatability is the second level.

Uh the third level is or is being able to use multiple tools. So the problem with custom GPT is still the data is stale. Um and so let's say you're using custom GPT to update your resume for every new job that you're applying for. You want to tailor your resume to that job and you want you'll be usually copy pasting stuff if you're just using custom GPD. So integrating with other tools like I right now again all the um all the chat bots give you connectors to directly link your Google drive or if you're more sophisticated sophisticated you could use MCPS or other ways to basically bring data from another platform here change the data and send it back to that platform. I think that's the third level of tinkering around that one can get familiar with. And then the fourth is the agent thing that we are talking

about where you're giving more complex tasks of decision- making nature. And again, depending on your comfort level, you don't have to let your agent, you know, message someone and make that call on your behalf, but you'll still be that human in the loop, so to speak. But I think that's the fourth level. So um I think when people think about learning AI, everyone thinks about the agent agentic use cases and think, "Oh man, that's so difficult. I'm not there yet."

And don't even like try the first kind of few steps to get there. So I think that's why setting aside the time and hands-on with level one and level two will quickly get you comfortable to go to level three and level four. I love that. And I just the only thing I would tack on to what you just said which is pretty much perfect is that and you kind of already did actually but I just want to highlight that the way to learn these tools for me is not as checking a box as for a new project on your resume and such. But unless it is a personal problem that you have to figure out for yourself I just find that people just don't go follow through with it.

you know, it just becomes like a thing that, oh, I have to make this project. I'll get to it. And they never do cuz it's just not something that they have to do, right? So, what I tell people is figure out the one thing that's been bugging you. There has to be something here, right? Like, I can't take You can't be like, "No, there's nothing. I'm perfect. Sorry, cuz then I don't know what to tell you." But find that thing out and then just fix that.

Just go talk to AI. Just blab for like 5 minutes. just dump all of your word vomit into the thing and I just tell them that tell me how to solve this tell me how to build this whatever you have right I think for me the biggest unlock that I have faced is I don't remember watching a YouTube video or a tutorial over the past year cuz I have the greatest coach like in my pocket so why would I watch a video that's doing like a thing parallel to the thing that I want to do when I can have somebody literally tell me step a 1 A 1 B 1 C, right? So, yeah, just kind of wanted to throw that in there. And I guess yeah, for our I guess like final segment here, I'm curious when you were back on the job market after, you know, such a long time. What was that like for you? Like was the hiring situation like pretty stressful? Were

you like pretty laid-back kind of can you talk us through your own job search that you kind of just finished recently? Yeah, sure. Uh so when I left Google I didn't exactly know what I was going to do. Um I I just knew uh two three attributes that I was looking for which is again I wanted to work in AI. Um I had again previously been in the finance and strategy team at YouTube. I knew I didn't want to do that exact job again. Uh because the reason for leaving Google was to stretch myself and learn something new. Um and so uh it was AI, it was uh having to do something that was more customer centric. Um and then the third was uh I wanted to be in a place where I uh could have more ownership and kind of see things end to end and see the output of what I was doing. So I didn't want to work in some place that had a long cycle from when I

started working to when I'll actually start seeing the output of things. So that was the criteria and in full transparency I explored uh it was kind of my period of what do I want to be when I grow up and I explored everything from do I want to become a content creator to you know uh do I want to start my own business and uh it was a again a period of exploring all those whatifs um that I wanted to do. Um to answer your specific question about jobs, um I I uh actually I I left Google and two months within that um a friend referred me at a company I was very excited about and I was like this is the only job I want and if I get this job I'll be happy and then you know I didn't get that job and so not getting that job for me felt a little bit of uh that anxiety around hey there are layoffs happening and you know this one not

getting this one job felt like the uh wow this is going to be a very tough market but again I got very lucky with uh landing at Ariia uh one of the advice I would have for anyone in the job market is to network extensively uh I didn't do it while I was at Google and um the the period that I left Google that's all I was doing I was networking a lot I was being more transparent about what I was looking for and also So um that allowed me to also um maybe self-reject some jobs that are not what I want or not where I wouldn't be good at or environments I wouldn't like. Um but I think because I was networking without an agenda, people were also a little bit more open to introducing me to others who are hiring for what I was looking for etc. And so that's how I landed my job. So I think that uh I think is even more prominent now because

of the number of AI résumés people are inundated with. The personal connection goes a long way. Yeah, it's one of those things, right? Like the best time to network is when you don't have any reason to network and I mean that that's what kind of makes it so tricky. Um at your current role, Ashwaria, what is something that you're really excited about? I'm assuming this lets you do the thing that you wanted to you know which is control and entire process slash cycle slashoutput end to end but in terms of just your day in the life uh or maybe you know when you look ahead like 6 months or a year what are I guess some of the things that make you feel really excited about the work that you're doing? Yeah. Uh so you hit the right bullets.

Uh so in addition to some of those kind of professional expectations I had another very important criteria for me was also the people that I work with and the mentors I would have uh at my work and I'm very excited about the team that I'm working with right now. Um so there is one the I'm excited about the stretch opportunity but also being with mentors who will challenge me and also kind of uh support. Uh and then I think I work in a company in the travel industry with a lot of legacy processes and I think AI again is that huge opportunity to just rethink about hey things were built a certain way historically because we again had a very different way of thinking about things but now with AI how can we untangle some of this and how do you think about what your infrastructure looks like for the next

10 years um and I think so That's a very again energizing problem because it kind of maps back to some of my formative experience at Google too where again I had joined some of that machine learning uh initiative and AI was still or sorry machine learning was still new things were not very formal the tools were not very mature and kind of getting that experience to build things from the ground up when everyone around you is learning and building at the same time is a very cool place to be. It feels like a massive college to be part of because everyone's learning and then you know when you fail it's not like a big failure because you're also learning from that failure. So I think that's kind of been what's uh interesting at my current role. That's so cool. Yeah, I can tell how energizing it is just cuz like just the

way you talk about it, it really felt like yeah, this is something that you've really bought into. You're locked in. you're like ready to disrupt this this entire industry as as it is right now. So yeah, really appreciate you sharing that, Asha. This has been so incredible. It's not every day that I get to pick the brain of a literal industry veteran such as yourself and I really want to appreciate you and you know thank you for taking the time to join us on the show and share your background, your journey, your learnings with those of us that are just a couple of steps behind.

So, thank you so so much for taking the time today. No man, thank you so much for having me. This has been a terrific conversation and you've been a very gracious host.