The AI Playbook Used By Global Companies (ft Olga Beregovaya) — Tool Use - AI Conversations

Let's be real here. We're not even in AI 1.0. We are in AI 2.0. And 2.0, it's where it's omnipresent.

Welcome to episode 48 of Tool Use, the weekly conversation about AI tools and strategies. I'm Mike Bird, and today we're talking about AI tools to save time and money while going global. When working at a global scale, being hyper-efficient can make or break a company, and we're going to explore which AI workflows can reduce costs for customers by up to 80%. This week we're joined by Olga Beregovaya, the VP of AI at Smartling, a global leader in translation. Olga, welcome to Tool Use.

Thanks so much for having me.

Absolutely. So do you mind giving us a little bit of your background?

I've been in language technology, localization, globalization, internationalization, natural language processing industry for over 25 years. I come from a structural linguistics background and then pretty much needed to take classes continuously and learn on the job. As it was becoming more and more obvious that structural linguistics by itself, as exciting of a field as it is, might not necessarily be as applicable when it comes to real job market. So there I am getting out of grad school with my second master's degree in structural linguistics, like, okay, what am I gonna do with this? And then it became very clear that NLP is the path to go. And then obviously if you look at the evolution of NLP, you go, okay, natural language processing, a little bit of computing. And then as it evolves, you inevitably find yourself in the world of machine learning, aka AI. So that's my trajectory and that's my journey.

Actually very cool. So being like from the origins with NLP, getting into more of the wide-scale LLMs, how have you found integrating AI into your, your personal life or your work life has kind of gone throughout the past few years?

I think the first time where it was like, okay, things are changing dramatically was probably when the tooling natural language processing tooling and overall globalization tooling started pivoting from rule-based to more potentially statistical-based and then machine learning-based. So you could just see that the tooling itself, the open source libraries that are out there, like then of course Hugging Face is taking over the world, right? So eventually you just realize that, okay, good old tools still serve the purpose. But professionally there is so much more that we can do with actual modeling and letting models making decisions. So I would say professionally, probably when GPT-2 came out and I started playing with it, I was like, holy Lord. One was okay, like, okay, yeah, no, no, we can live with that. But then when two came out, it was very, very apparent that the world of globalization and translation is never going to be the same. So it's basically gone from, hmm, interesting. Well, oh wow, scary to okay, here we are, inevitable. So I think now we are just listening to Andrew Ng talking at a conference who said, look guys, I mean, let's be real here. We are not even in AI 1.0. We are in AI 2.0. And 2.0 is where it's omnipresent. So I think, yeah, I mean, equally professionally and obviously our personal lives are, we get Netflix recommendations, right? We get like, we get recommendations everywhere, everywhere we go. Sometimes I'm laughing 'cause sometimes I have some questions to the UX of AI assistants, uh, assistants, because sometimes I'm just thinking like, okay, I'm sorry, it might sound horrible. I just want to be honest. Sometimes in the personal life and in my own productivity world, am I back in the world of Microsoft Clippy? Like, I mean, this thing just pops up and basically takes over the process. I think UX-wise, there are things to do, but otherwise we're definitely there.

Do you have any thoughts on that, how the UX is going to evolve? Because everyone is— we're getting tired of the chat interface. There's always that empty box that you type instructions into to start from scratch, and there's so much potential how it evolved. Do you have any predictions on how things might go, or just even in your idealistic world where you'd like it to go?

Yes, I think, and that's where we, I think we start talking about, um, agentic workflows, right? So one thing is part of the UX will just natively disappear because part of the UX will just inherently be taken over by agentic workflows. So, you know, you set the initial parameters, you go with your asks, and then you really, uh, let the agentic environment, whether it's agentic with a capital A or agentic comprised of multiple agents, eventually that environment will remove some of the UX. Now, my prediction for the more copilot assistant-based UX, I think that eventually we're just figuring it out. And I think eventually it'll just be a little bit less intrusive and a little bit less invasive. Because right now, sometimes I have a feeling like, oh damn, let's just plug the chatbot somewhere and let's say like, hey, I'm your assistant. Like, hey assistant, I was going to write this letter myself. So I think it'll be more subject to user configurations and user preferences. And I also think that, I think just the arrangement of the real estate, whether it's in English or any other language, I think the arrangement of the real estate would just be a tad less invasive. And we can talk about UX for, I mean, Now we need so much UX. We need, especially in our industry, you need prompt work surface, you need post-editing work surface, you need fact-checking. I think the whole world of UX translation on translation will never be the same.

Yeah, I fully agree. It's going to be a lot more ambient. It'll do things for you before you ask based on other triggers or whatnot. It's very exciting. As a background in systems integrations, just knowing that we can kind of plop a little AI right in the middle of a pipeline and you can get more functionality out of it. I'd love to pivot to Smartling. How does you or how have you approached incorporating it into whether it's workflows or at the engineering level? How has a global company adopted AI?

I mean, we are an AI-enabled company, right? And we offer and provide, I'd say state of the art, actually. I'd say, yeah, definitely state of the art AI-powered AI solutions to our customers. So when we tell our customers like, hey, AI is really what's gonna help you gain efficiencies and cut costs, we better be able to eat our own dog food, right? We cannot just, I mean, practice what you preach, right? And that's, I think that's a necessity. So we use, let's start with the translation process. We are equally services and technology company or AI-powered full end-to-end services solution. So we provide the best experience for our translators, giving them productivity tools, using automated quality estimation, using automated quality assessment, using agents. So by the time our translators, and I'm talking about our internal translators, by the time the translation workflow hits their desk, a lot of preprocessing, AI preprocessing steps have already taken place. Then, interestingly, back to the UX. There is something in our industry that I suspect you might not have heard of, which is called CAT. Not a cat with a tail, but a computer-assisted translation. There are so many industry memes where it is a cat, but it's not. It is computer-assisted translation tool, and obviously we keep on enhancing our computer-assisted translation tool. To plug in more AI. Like right now, for instance, translations get ranking of potential sentence complexity to see how much cognitive effort they actually need to put into that sentence. So the string— translator is shown string and translator is shown string with a label like, hey, works, doesn't work, good, bad, ugly, indifferent. So that's for translators. Now we have the most amazing transformation in our lead generation and business development function. And right now the effort has been cut down, uh, I don't remember the exact number, but let's say by 60% 'cause there is so much you can do with lead gen, right? There is so much you could do with targeted mass mailers. You can do so much with tracking how your leads actually, whether they've been responded to or not. So I would say that our business development function was probably one of the first adopters in the industry, in, and definitely in our company where we actually work with our global potential prospect base using AI. So you already have two, you have translation, then you have lead generation. Then obviously my team, my team of data scientists, do you have any choice but using AI? So obviously I'm not even gonna go there. I would say the late, I mean, obviously all the modeling tools, quality estimation tools, LLM as a judge, all of that is natively a part of my team's workflow. What we did, what we introduced recently is we really looked very closely at DSPY, right? And we are pivoting from prompt engineering to prompt programming, right? I mean, and I think that's where the world is going. Again, our industry or not our industry, again, it's there are only this many prompts a human can engineer, right? But if you give enough examples and if you are, I mean, if you're very explicit with what you want out of that prompt, prompts, writing prompts is a part of my team's workload. And then obviously you have the whole programming and our engineering team and they're obviously, again, that's where the world is going. The world also is very well aware that AI writes code. Now, is this code always perfect? Question mark. But are we using— I mean, but our engineering team is very diligent at using AI and testing AI before anything is implemented. So here you are with pseudo-testing, right? LQA automation. What else would there be? Coding boilerplates using Git Copilot. So I would say that we internally are very AI-powered organization. There is still work to be done. We're actually running an internal hackathon where we're gonna be, we polled every function in the company and said, okay, where, where, where are you, uh, where are your efficiencies? Where can your efficiencies be improved? And now we are going on a 90-day hackathon to see, okay, what else internally? So that's again, eating our own dog food. That's, uh, that's what we do.

I, I love that. I'm a huge fan of internal hackathons, giving people who are actually like at the ground level doing the operations, the ability to kind of take some leeway to try to build tools for themselves or whatnot. Very powerful. I'm curious how you go about the build versus buy debate where there's a few off-the-shelf tools every day. There seems to be more and more, but a lot of companies with the resources will build their own bespoke tools to accomplish certain goals. Do you have a process for going through either auditing what exists versus just building yourself, or are you partial to just, you know, let's just build it because we know our stack better than anyone else?

I would say the process, I mean, first and foremost, there is money, right? I mean, you look at the ROI and you sometimes work backwards from the ROI and you see, okay, I've just invested into this, let's say model testing tool, models management and prompt management and prompt testing tool. And you are like, okay, here I am, $100 grand down, but I have a sneaky suspicion that they might be using the same approaches, potentially the same software libraries, and potentially even same, like as I said, models as a judge. And then it was like, okay, well, here we go. We can actually build it. And that's, for instance, how we build our model performance estimation tool. Yes, there are tools in the market, but quite often, and I think we also need to talk about the democratization of AI, everybody can access it. And we also happen to be AI experts. So when we see that, okay, commercial ROI is not there, and we can build it ourselves, then we would just go with a build. So I think ROI and do we have the talent? Are we willing to, like, you measure the t-shirt size, are we actually willing to go in there? Shall we do it ourselves? And here we are, our model performance management suite uses the standard metrics. We see commercial solutions and we can most definitely build it ourselves. So ROI and capabilities. And I would say, when would we make the buy decision? I think it's again, looking at internal capabilities and just looking at, okay, is the juice worth the squeeze? Is there something that's reasonably priced? Is there something that's commercial? I mean, is it something that, like for instance, is pay-as-you-go or unlimited subscription that would take us light— that would take us years to build? And that's where we would make the buy decision. For instance, our sales process and business development process optimization, we definitely were very happy with off-the-shelf tools. So yeah, really capabilities, come in, come into it with open minds and also be very, very realistic on shall we bother? And if we will, why?

Always got to have a justification for it. Along the same lines, I'm thinking, how do you know which experiment to run first or which process to prioritize first? Is it the justified by the bottleneck in the business, or do you wait till a certain engineer brings up a good use case? What's the process for saying, okay, here's the priority list, let's run these experiments first?

My hypothesis, and I think it's a very well-validated hypothesis, is that innovation consists of 3 pieces. Piece 1, state of the art. What's out there, right? What's best in class? And we all live in the same universe, and that's, that's one, uh, one universe where, uh, You wake up to 5 new models a day, right? Or 10 new models a day or versions of the same model. Like, hey, you know, we just added a couple trillion parameters. Well, okay, thank you. We just finished our experiments with the previous version. How fun is that, right? So part of it is what's state of the art, what's out there, and monitor very closely both academic research, commercial releases, and open source releases. I mean, we all read like, I don't know, 40 TL;DRs and 45 newsletters and follow again religiously what's happening in Hugging Face and other communities. That's a third, right? One-third is just our own vision of where we want to be as a product and where we want to be as an organization. And that would drive, okay, these are the experiments that we need to run to cater to this product vision, right? Like, for instance, we decide to introduce for internal and external purposes automated language quality estimation, language quality assessment. And LLMs are not, again, they're not very bright when it comes to assessing language quality. So that's where you're like, yeah, hey, we need it. It's a pressing need. You can look at 5% of your content or you can look at your content across the board. So business requirements would be, um, business requirements and vision would be one third and the other third would be just listen to your customers. The customer comes to you, the customer comes to you with a problem, and that's where we, we would know what experiment to run. Now back to my initial statement about not to fall for the frenzy. And not to be enamored with like everything like, oh damn, that's a new, that's what I've gotta experiment with. We have very robust testing and benchmarking environment. So before we embark on any kind of experiment, we just assess the model and we see, is the model going to do what we want it to do? Like for instance, our agentic functionality, we would usually know, is it worth our while to look at something new or are we good where we are? So, and then just again, the market, traditionally translation market was dramatically text-based, and now you like it, you dislike it, you need to experiment with multimodality. So it is also where the world is going outside of our maybe a little bit of myopic localization space.

Hope I answered your question. Yeah, absolutely. And also, I'd love to double-click on the benchmarking aspect. So it's very important to companies to have proper eval setups so that when they introduce a new model or try something new, it's not going to cause a regression. Do you have any strategies for how companies of any size can implement benchmarking to make sure that they, when they do an upgrade, it's actually an upgrade and not just, oh, it worked well in case A and then every other case is regressed?

Okay, here is our world, our world of global content delivery. There is, what, 1,732, unless I'm mistaken, world languages. So every model change, process change, pipeline change, just multiply it by 1,732, and then you'll find yourself in our universe. And we quite often find ourselves in, I mean, we're blessed with having multilingual workforce. So we have people from different backgrounds. It's very easy to pick up the phone and like say, hey, such and such, do you mind looking at Spanish and Portuguese? Or like, I mean, I speak a few languages, so we can do that as well. Our engineering team is pretty predominantly Eastern European, so we have those languages covered, but I guess my message is we need, as a global content, AI-powered global content delivery company, we need to think global. We have no choice but. So for us, we split languages into language complexity groups. We by now have accumulated enough knowledge about, okay, Brazilian Portuguese is going to be a dream. For this set of, for this set of LLM, whatever the assessment, translation, or post-automated post-editing or anything in between, but don't nurse the same hopes for Turkish. So when we test models, when we benchmark models, we absolutely benchmark across different language groups and different content types. We actually are super rigid about our golden dataset, In collaboration with our quality assessment team and data scientists and engineering, we develop a golden dataset across different language groups, different content types, different string formalism, basically golden dataset that would allow you to run a suite of tests. And once you've run it, you usually have a pretty good idea, okay, this is where I'm in danger. Now, it's very important, again, I said that sometimes we ask a human to have a look. We're also religious about direct assessment and engaging humans to give us a point of reference, but quite often it's a press of a button. And here you are with your F-measures, right? Precision, recall, accuracy. So it's somewhere between having a super robust golden dataset and benchmarking environment that, like, you press a button, you look at a table, and you see, okay, danger, no danger, red flag, let's proceed with further experimentation.

A few people have mentioned the golden dataset, and that's becoming the norm where you just need to know what data you're actually looking to get out of the system when you put a given input in. So when you make a change, things are looking good, all the evals are passing, what's the deployment like? Do you try to do a gradual rollout or is it more of flicking a switch? At the global scale, how does one deploy AI into their systems?

Okay, well again, in our specific case, I run the R&D department and we work super closely with our product and engineering department. So whatever works, and it's what I'm going to say is absolutely mission critical. It works in the lab. You start deploying it at enterprise scale. And we're like, I mean, we're a huge enterprise scale platform, right? I mean, we have clients from all ways of life and clients of all sizes. You can see all the huge logos on our website. So here we are, ta-da, things work. Yeah, yeah, go global, 5-minute latency, goodbye. Right, or like inference taking an hour. So we always make sure, and any organization looking to deploy a successful AI implementation into their pipeline absolutely needs to run latency and inference tests and make sure that everything that runs in the lab actually is scalable, have a fallback mechanism, and just ensure, you know, whatever there will be queuing or multithreading, multiple techniques, LLMs still lag. And again, we all know that, right? Some have mitigated it, some have not mitigated it. We're blessed to be able to work with data science teams from the most prominent research companies out there. But first things first, check for latency, run your pseudo-tests in the sandbox environment, and make sure that either you don't introduce any delays or your engineering, your brilliant engineering team has developed techniques for mitigating potential issues and deploying at scale. So that'd be number one, making sure don't rest on the laurels if something works in the lab, right? So first rule of deployment, basically make sure that it's not gonna introduce any surprises and always have a plan B, always have a fallback mechanism. So that'd be that. But there are so many other deployment things you want to take into consideration. Again, things work great. How do you know whether the model is hallucinating or not? So you obviously need the guardrails for mitigation of potential semantic issues, right? And that would be, is the model dreaming things up? So you absolutely need to have hallucination mitigation mechanism. And I'm not quite sure we're there yet with self-policing models. So you do need some form of hallucination mitigation. So you have scalability, you have hallucination. Now throw it all in or deploy gradually. I mean, if you have a specific feature in your product or specific efficiency potential in your internal processes, you probably want to deploy that entire feature, right? And test that inside there. Because I mean, what good is it going to do us if we just deployed our adaptive translation memory, right, and did not see how it works in accord with our, for instance, glossary management and glossary insertion. So, I mean, services architecture, deploying by feature, but definitely not throw the whole thing in. I would say deploy gradually. We're an agile company, so once the tests have been completed, usually we operate in 2-week sprints. And usually when, again, research has come up with something, product has vetted it, usually engineering takes 2 weeks to deploy it and then test it. So continuous, nothing, nothing, maybe nothing new, right? Continuous development, continuous integration, continuous deployment, but risks, right? Mitigate your risks, know your risks and mitigate your risks. [Speaker:OLGA Beregovaya] Exactly.

We get enough risks by introducing deterministic code. As soon as you bring LLMs into the mix, it's just, you know, always have a plan B, like you said. One thing I'd like to touch on briefly, you mentioned hallucination mitigation. Do you find LM as a judge is sufficient for that, or is it still going to require human in the loop for something as important as translation to get right?

I'd say that it all depends on the content risk tolerance. And again, we are, I mean, we are all about, again, we don't fall for, we don't fall for the wild, wild west of AI. We're all about measured. Deployment, and we have a fairly strict and fairly rigid grid of what content we can send through AI-only pipeline and live happily ever after, and what content types, for instance, we work with regulated industries, we absolutely would say that human would need to have a look and fact-check and check for hallucinations and anything that, I mean, the usual suspects, hallucinations, biases, You decide based on the content type, based on the language complexity and content risk tolerance, you decide how much of human in the loop you actually need. And there are a lot of times where you actually do need human in the loop. Now, no, again, back to 17,032, talking about fact-checking, I'd like to check my facts on how many world languages there actually are. But back to that, we all do know that models tend to hallucinate much more in other languages. And a lot of implementations outside of globalization industry are quite often either monolingual or naive. Monolingual, right, is when development and product is English-centric, and naive is, oh wow, OpenAI API, ChatGPT API, Gemini, you know, Vertex AI API, here I go, I solved all the world's problems. And I think we are again through testing, through correlating with the human things, correlating with direct assessment, we usually have a pretty good idea of how models are going to behave in a particular language. Now, having said that, we do have automated AI-based checks in place that solve probably for 90% of hallucinations, of model hallucinations, but then it's your own decision how much of the human you want to, based on the parameters above. You want to, you want to see. But LLM is a judge and especially even LLM, even more so perhaps one LLM judging the behavior of another LLM. It works in a lot of cases.

One thing just along the lines of eval, you mentioned earlier DSPy and I'm a fan of DSPy. I fully agree that the idea of prompt programming is going to become more important as people kind of figure out what patterns really work and what structure gets better results. How do you work with evaluating the output of DSPy where the normal process would be, you know, you have a given input, your prompt, and a given output, and then you just rank it. But DSPy, it's an extra component in the middle that adds a little more twist to the pipeline. Do you have any strategies around evaluating how DSPy is improving performance in a system?

I don't necessarily think that we're looking at DSPy for improving performance. I think we're looking at it more as an efficiency tool, right? And just basically sparing the time as opposed to fully human, full human prompt engineering. Now, again, we're in a very fragile space. Linguistics, language is a fragile thing, right? Human languages, you really want to make sure that you deliver accurate outputs even more so when it's translation. So as of right now, I would say, and let's also remember I'm not the legs on the ground or feet on the ground, whatever it's called. I wish my colleague who works predominantly with DSPi would be here. but again, we usually just look at the accuracy of the output, right? We program those prompts for a specific, for a specific task, and usually that's where you would measure whether the task was performed or not performed compared to human-engineered prompt. And this is where again, you have a gamut of automated traditional NLP metrics, both text-based and semantic, referential and non-referential metrics. so you can apply those to the output. And also, again, we make a fairly strong investment into labeling and see how the DSPY output actually delivered on the task. So somewhere again between human direct assessment, applying automated metrics, I would say that we measure the performance based on the output and based, based on whether the task was delivered on, was executed on or not executed on.

So Smartling definitely sounds like a company, like you said, AI native, AI has been incorporated into a lot of different aspects of the business. When you talk to your colleagues or, or just see different companies operating, what do you think slows them down from integrating AI into all these workflows? Why, why do some companies adopt it very quickly and are willing to put it into the processes that matter while others are kind of dragging their feet and potentially getting left behind?

If we look at enterprise AI deployments, there is one factor that makes companies, uh, move very quickly for the good or for the bad. And it's C-suite mandate to implement AI. And that's, I was at a conference, like a, or even more so probably private event not so long ago, which resembled a little bit of support group. We were just mandated to implement AI. What do we do? How do we implement it? And sometimes again, for the good or for the bad, without necessarily measured deployment, a lot of companies are just forced into implementing AI into their workflows because they need to, because they have to, because everybody does. And as we know, I don't know, it's probably a year-old data, right? That 75% of AI implementations are not successful and 25% of AI implementations are successful. I guess that's the data from a year ago. I'm pretty sure those numbers have changed. I think a lot of companies rush into it, but then 7 experiments did not work, 7 implementations did not work, but 3 did. So sometimes it's just a mandate and it's like, I have to do it because I have to do it. Some companies, I think just more of the innovative and adventurous spirit and maybe a little bit less risk adverse. Some companies, I would say that just the structure of their product, if they implement AI features into their product, just the architecture of their product allows them or architecture of their platform allows them to easier integrate additional components, right, into their less monolithic and more service-based architecture. So I think it also has a lot to do with what the platform architecture looks like. So mandate, less risk adverse, that platform and the product just allow them, or whatever that is, a solution, allow them to integrate additional AI-based components because it's easier to do so. And I would say the companies that are slower, I would a little bit divide it by verticals. Again, I mentioned regulated industries and regulated industries just traditionally are more conservative when it comes to implementing new technology. I lived it with machine translation. I lived it with text summarization, sentiment analysis, and now AI. I would imagine that regulated industries like life sciences, I just slower at implementing AI and probably for yours and my benefit, right? I mean, they would probably, they probably should be implementing patient chatbots or like adverse events analysis after thorough testing. So yeah, I think, I think these would probably be the factors. And sometimes the company would move slowly, but they would go through very rigid compliance regulatory, ethical AI process, they can move slower, but they will probably land at a safer and more measured place. Excellent.

Um, so when I first heard about, uh, Smartling, I thought translation is one of those areas that seems ripe for AI to just completely decimate where everyone kind of has a translator in their pocket. But the impression I've gotten, feel free to correct me if I'm wrong, is that Smartling's not going for the general, like everyday translation. It's going for mission critical when you need an accurate representation of text in another language. How do you foresee different industries which are at the threat of AI surviving? Do they have to go along the same lines of work up the chain? So instead of just going for mass market satisfaction, which AI can probably solve in its own realm, but something where it's either more valuable or more, um, just a, a higher need for success. How do you see other industries kind of surviving this AI wave?

Well, we always know that you definitely want to be a part of the solution, right? Rather than be a victim, victim of the solution. So I think that by now, through trial and error, and again, that's my thinking, '23 was, oh, you know, oh damn, AI, hey, well, we don't know what to do with it. '24 was probably a year of learnings and trial and error, right? And implementations. I think by now, here we are halfway through '25, we're at a place where pretty much narrowed down the use cases where it actually does work, right? And we're talking about entry-level coding, right? We're talking about, again, forgive me, people in those industries, but we're talking about technical writing. We're talking about SEO. For all we know, SEO as we know it may be a thing of history, right? So there are so many industries, verticals, trades that most obviously are impacted by AI because we do know that by now generative model, I mean, let's leave AGI out of the equation for now, but generative models, other ML models do deliver. And I just mentioned a few industries and translation industry is one of them. I think understand where it works, understand where it cuts out the necessity for human steps. Understand and reinvent the role of human professionals is basically where the industries maybe similar to ours that work with text, that work with image generation, multimedia, voice generation. I think just understand, okay, this piece is now safely taken by AI. How do I reinvent myself and how do I reinvent the workforce? As I said, inevitably there'll be jobs that will be gone. In our industry, there is this ongoing debate, do I want my child to be trained as a translator? And quite often the answer is straightforward. Translator is a no, but do I want my child to be a multilingual prompt engineer? Hell yeah. Do I want my child to learn to be a program manager and analyzing business intelligence delivered by AI? Yet again, hell yeah. So we talk a lot. I personally talk a lot and mentor students on, I wouldn't go as far as survival techniques, but there are so many roles for you. And even if the industry seems to be at threat, beef up your skillset and you'll be thriving.

I'd love to actually get into that, that granular, the human level where we are entering an age where everything's changing faster than it ever has before. What skills should people even focus on? 'Cause I love the way you framed translation versus multilingual prompt engineer. What other type of pivots or slight adjustments should people take to their education or their reskilling if they feel threatened by the oncoming AI wave? What, what do you feel is a valuable skill in the next few years?

Okay, so here is my own son. I hope, I don't know, probably, I don't know if he'll listen to this or not. My own son is getting, uh, his master's in, uh, data science. But being a, well, first of all, mother, second, control freak, and third, industry professional, I check his curriculum religiously, making sure that they are actually being taught what's not only relevant today, but will be relevant in a year, 3, 5. But no matter, I mean, yes, and I'm fully satisfied with their curriculum. But that aside, I think that familiarity even Basic familiarity with data science is an absolute inevitable. Understand statistics, understand the basics of machine learning and data science, no matter which field you are in. Either you'll find yourself directly working in some form of data science field, even in the rudimentary form, or you will be furnished the output of AI and you better be able to interpret it. And not just say, oh damn, the model just spat something out. I have no idea where it's coming from. I will again go back to translation industry. The most sought-after unicorns are somebody who is multilingual, understands the principles of linguistics, can test and iterate whether it's prompts, whether it's RAG, whether it's whatever, wherever the content is coming from, and most importantly, interpret the metrics that they receive. Initially, we started with a very naive and again, I mentioned data labeling before. Initially, we started with thumbs up, thumbs down. It's good enough, right? Okay, well, it's formal in German, it's not formal in German, happy, unhappy, thumbs up, thumbs down, but it only gets you this far. But if there's a unicorn out there who equally understands the core principles of how the models operate, how, how ML operates and how LLMs operate and are able to interpret the signals. That's an absolute precious skill. So my forecast is get a degree in the trade where your heart is at. Wherever your heart is, pursue your dream. You don't need, I was just speaking at a linguistic conference where people are like, 'Hey, we're in humanitarian field. We're not— humanities, humanities. We're in humanities field. We're not born data scientists. What do we do?' Take yourself where your heart takes you, but make sure that you augment your potential trade with ability to understand the core principles of everything that leads to machine learning and machine learning and data science and be able to apply to your day-to-day. Yeah, again, quoting, just presenting at, attending and presenting at that MIT conference and quoting the, quoting Andrew Ng again, you better know how to program even a little bit, even a little bit because you will, again, entry-level programming is gonna be a thing of the past, but you understanding where the outputs are coming from are essential. So do what you want, augmented it with additional set of skills and you're golden.

I think that's wonderful advice. People with a bio degree or life sciences, where they feel they're completely in the physical world. But if they just understand AI literacy, like you said, both the interact with AI as well as interpreting the output, being skeptical healthily is I think very important. So that's wonderful advice. Last question for me, in your day-to-day life, what AI tools do you use and what do you wish existed to help make your life just a little bit easier?

I mean, obviously the search, all of my search has migrated to Perplexity. ChatGPT. I still do Gemini summarization and I usually have, I think, I think at this point in time I'm pretty, I think I pretty well understand which of the models will deliver the best results for a specific task. So I would say search and analysis most definitely. I am still very, I was talking to this professor who's like, why am I even teaching people when they can write their PhD? They have, they can have an LLM write a pretty pretty sufficient PhD, uh, within 5 minutes. Why bother? But I still like to do my own writing, maybe just to overcome that staring at the blank page, the writer's block. I definitely, when I write articles or when I write blog posts, I do, but I always just use it as the bare bones and as boilerplate. And then, and then obviously, moi-même, myself, I was talking to a recruiter who said that she received 17 identical CVs. And she was not thrilled. I just want to make sure that I don't find myself, find myself in that space. One piece of advice, again, when you beef up your career, there is so much happening in the world of natural language processing, um, GenAI, AI in general. There are so many articles out there, but who has the time? Just use summarization. Just use summarization, extract core concepts, and here you are, you've gone through 10 articles over one evening. So I definitely do a lot of that. Absolutely do a lot of data extraction and a lot of summarization. And then again, I was speaking to Professor of Ethics who said that there is a new field in ethical AI, which is treating AI addiction. It's becoming a thing. And when I caught myself, when I caught myself trying to write a poem about cyclists and motorcyclists, in the style of Homer's Iliad. I was like, okay, well that's a fun way of spending the evening. So I'd say for entertainment purposes only, just to see what that thing can do equally in images and text. And obviously machine translation, that's not going anywhere. So I'd say it's omnipresent, it's everywhere. I just wish UX was a little bit less intrusive.

[Speaker:DAVID] Yeah, and, and I think we'll get there. And actually just on the blank page problem, one thing I've kind of grown into is I actually have the AI generate a list of questions for me to answer to adhere to a certain structure. So that way it's not actually writing anything, but it's prompting me along the way so I can kind of keep things targeted and structured in a more productive way. Because I agree, like thinking is paramount to writing and we need to keep doing it. We can't just offload everything to the AI. But Olga, this was awesome. Thank you so much for coming on and sharing your insights. Before I let you go, is there anything you want the audience to know?

Anywhere they can keep up with you? I would say that the best thing, I mean, first of all, I have my own LinkedIn page, so feel free to add me on LinkedIn. And usually the interviews, the articles would always be linked to Just add me. But even more so, Smartling is super active on LinkedIn. So if you just add Smartling, if you just follow Smartling on LinkedIn, you'd have pretty good visibility into my team's work and other teams' work. If you go to Smartling's website, we, again, we post whenever there is a podcast, whenever there is a webinar, whenever somebody writes something, it's always reasonably easy to to follow us.

Awesome. Well, thank you very much.

Okay, thanks so much for having me.