In this talk, Amber Roberts provides some insights for job seekers on finding your first role in tech. Geared especially to engineers and data scientists, Roberts takes questions from the audience about their own job searches, and gives some actionable advice. This talk was originally recorded at Arize:Observe in April 2023.
Amber Roberts: All right everyone. Thanks for joining. My name's Amber. I'm an ML growth lead here at Arize, and I'm gonna be walking you through kind of the key things that I've seen help folks get that next role in tech. I do have a blog post on the topics we’ll be going through and I'm gonna kinda walk through those and share.
But I really like people to share their own journeys. Um, I like these to be as collaborative as possible and I think when. All of us have tried to get, you know, our first role in tech, our first role, um, out of grad school. We've attended talks and there's always been kind of caveats of, oh, that doesn't relate to me.
So I wanna hear, um, from you folks, and I really wanna answer your questions. So my background, um, so I am an astronomer. And then after, um, I got my master's in astronomy. I applied for a lot of different roles as a data scientist. Wasn't really sure what I was doing. I was sending out, you know, 3-page CVs and not hearing back.
And then I joined a fellowship program at Insight Data Science and it was the AI track for that program. And then after completing that program, I worked for Insight for about two years and ended up being the head of AI at Insight and the key parts that we were doing is we were bringing in a lot of fellows, bringing in a lot of folks with different backgrounds, sometimes technical, sometimes non-technical, and helping train them to be data scientists, data engineers and machine learning engineers.
And throughout that process, I've really seen over 200 folks from different backgrounds, a variety of backgrounds, get roles as data scientists, as machine learning engineers, and as data engineers. And those key findings, I've been passing them along. I've had a lot of one-on-ones talking with new grads and just trying to help folks understand what the key things they need for getting that next role are.
So I'm going to share screen just to show the paper, but again, I like these to be super interactive. I like to really dive into different parts of what people are struggling with to give them realistic answers and to give them actionable, uh, answers. So being able to walk through, you know what, what the problem is, how do we get to a solution?
So in the piece I'm walking through, there's really five rules that I see. It being a numbers game, networking, folks saying it's not the right time, the change in environment and then also just noting like everything is an interview.
So before getting into this, and I'm sure there'll be questions on it, questions as we go through. I'm curious like, where, where are people at in their ML journey?
Are there any folks here that are looking for their first role, that are looking for their next role?
Any particular roles that you're interested in?
Um, and then if there's any challenges that you've. Met along the way there. Okay. And I think I should be able to just see the chat.
Cause I'm gonna be constantly looking back at the chat. So just getting into the main rules. So the first is, and everything we're gonna be covering here is from that first initial thought of, I want to get a role in tech. I wanna get a data science role, or a machine learning engineering role, to the point of actually starting that role. I recommend attending Lacey's session as well, cuz I will mention negotiation and knowing self worth, like she really will get into those topics. But if you have questions on negotiations, I can also answer those in there.
So the first one being it's a numbers game and it seems like that should be obvious to someone that's analytical, someone that's applying for roles in tech, but the truth is it's hard to kind of wrap your mind around: Hey, I'm playing the numbers game. Because you do get attached to these roles. There are ones you're very passionate about and ones that hurt if you get rejected or it doesn't go through.
Um, I also see we have a few people responding, a data scientist looking for an ML focused role, that's a very, very popular one. Thanks Calvin. Garrett. Thing he's looking for. His next role had previous tech and AI positions. Um, and I know there's a lot of folks now that are looking at roles like prompt engineer and um, LLMOps and these roles that are just starting to come out, but are essentially the new Data scientist and the new machine learning engineer.
So thinking first that it's a numbers game, uh, it really depends on the amount of applications that you're able to get out. Um, hearing from folks saying they're having trouble just getting interviews on the calendar, I tend to ask how many resumes they're sending out and normally the answer is like one or two a week, which just isn't enough. Um, remember, you don't have to prepare for these interviews when you're sending in a resume, you know, understand why you wanna work at the company, but I know some folks that will apply and they'll start even prepping before they hear any feedback.
So, make sure your applications are concise. You know, you can have, I would say a standard covered letter, but it doesn't need to be, um, super well researched unless it's a dream role. Um, you know, you can take a few things like from the website, the mission statement, see what things you align on, put those.
In your cover letter and make sure you put it in the job applications. I recently saw a post, uh, if I could find it afterwards, I'll link it. But how to use chatGPT to kind of tailor your resume. Um, you know, tailoring your resume per role has been done. Um, you know, since people started working in tech.
But I think the most common thing that people do is you see the job description first. See that you meet at least 50% of the criteria. I say if you meet at least 50% and it's one of those quick applications, and it'll take you maybe 15 minutes to submit an application or less, uh, just apply for it.
See how they phrase their requirements, like what they're trying to get from their requirements, and then retailer your resume for that. Uh, remember just one page. You know, no CVs. Um, and if you have something that fits more into a requirement, maybe just, you know, take off some of the competition, some of the extra projects, um, just to make sure that your resume is tailored.
A lot of these, you know, if we're applying for roles as a machine learning engineer and we know that our resume is likely gonna go through some kind of algorithm, we just, we gotta fine tune the algorithm. We know how to do this. Um, and so we should use our skills, uh, to do that and to give us the best possible outcome.
Uh, it looks like we got a question, um, about networking. That's a great question. So the next, um, So the next area we have is actually networking. So we just covered, you know, it's a numbers game, so give yourself the best advantage. Put in as many, like, just give yourself a goal. Maybe I'm applying, you know, three or four times a day, or like three or four applications a day, Monday through Friday.
You know, I would, I would personally say, do at least five applications a day, but don't spend more than an hour on a single application. That's not necessary.
So rule number two. Um, network, network, network. So sending out cold applications, cold applications where you have no connection at the company.
It's not gonna give you the best, um, ROI, return of investment for the amount of applications you send out. What you really want is that warm introduction where you know someone, um, at the company and yes, I know that's, well, that's ideal, but you can start networking. You today I've talked to a lot of folks who are PhD students and they're like, oh, well I'm gonna be in this role for the next five years.
Just start networking. There's so many folks that are just on this, um, that are just on this recording. You can connect with me on LinkedIn. You can connect with all the speakers, send them an invite, send them a message that you know you appreciated their talk. You'd like to connect, cuz those connections really build up over time.
So if you are able to just set. A goal for yourself is like, I'm gonna attend one in-person meetup a month, and I'm gonna attend four virtual meetups, um, every month. So like one a week. Uh, attending those sessions, making those connections, just really build up over time. And, um, the key thing, like with that network as well, is it just, it just keeps expanding.
So being able to connect with individuals, you'll see that while tech, you know, tech seems very large, you'll be seen the same people over and over again. And you'll be like, if you had a great experience with them, you'll want to work with them again. Um, they may want to work with you. You may even. You know, work together on a startup.
So there's a lot of options there, but just putting yourself out there, and that tends to be the hardest part for a lot of people. Like they'll, they'll crush coding interviews and they'll crush system designs, but putting themselves outside their comfort zone. Introducing themselves, um, asking people what they think of, you know, what do you think of the company you're at?
I'm considering applying. Doing that can be hard, um, especially if it's fully remote and fully virtual. But, you know, just do that. Um, and I even in my post put out a few kind of. Like just example wordings that you can use to connect with people on LinkedIn. Uh, for me, networking is primarily through LinkedIn.
Like if someone writes a great blog post, like I'll message them, um, you know, wanting to connect, like wanting to follow, uh, the work that they're doing, uh, meet someone at a conference like wanting to, to connect. And so just. Kind of making it personal, making it tailored, and just building up those connections.
And the more connections you build up, the more mutual connections you'll have. And especially if these are the roles that you're looking to get into, the more connections you have around that field, the better populated your LinkedIn will be with new roles. Um, you know, research coming out paper is like, I love going on LinkedIn because I have such great content to consume.
So just keeping in mind, building that network just really helps. You become the best machine learning engineer, data scientist, um, lops, uh, professional that you, you want so, Let's go into some of the questions here. Um, so I think I answered kind of the best ways to network. Um, you know, there are a lot of community events also, if you're looking for roles, there's different channels that you can look in.
Um, Chip’s MLOps, Discord channel for jobs in the MLOps Slack group, um, and Data Talks Club. So there's a lot of different slacks where everyone is kind of on the same page. There are different channels you can join and just messaging people through Slack and Discord is is a great way to connect.
Question from Calvin. So one of the issues that you've been experiencing recently is sending out a lot of applications. Okay. Getting some interviews. That's great. Um, already getting some interviews back. Um, but there aren't as many companies hiring as before. Um, And then you're, you're probably applying for jobs that might be continuously listed and maybe they don't take them off.
Uh, so any thoughts on the current job market? Wow. That like perfectly kind of segues into the next topic. So thank you for that one, Calvin. Um, if it's never the right time, it's always the right time, so. I've, I've heard over and over, um, pretty much this year, um, like we've seen, we've seen a lot of tech layoffs.
But if you look back to essentially since the dot com bubble bursting, there's been, um, you know, never like three or four great consecutive years to get a job in tech and when people say like, oh, it's just, it's the wrong time to graduate or the wrong time to look for the next role. You kind of just have to look back and see, well, the pandemic and then recessions and then, um, you know, just different parts throughout.
And obviously this is tailored to the US economy, but just looking back through, Um, different markets and you can check the economy for, um, your own country, but it's gonna probably look similar. The only thing that is consistent though, is the growth of jobs in tech year over year, um, and the amount of people hired for emerging technology.
So that's what tends to be the thing that separates, um, tech, is that it's constantly. Turning over, but it's more just pivoting. Um, so the pivot comes from tech being so new. Like I myself have had, um, if you look at my LinkedIn, And the roles that I've had, I've had such a variety of rules in the past, like four or five years.
Um, and that's also what really made me want to be in tech. Um, coming from academia, uh, folks here that are in the chat, coming from academia, you've probably experienced this too, where if you're coming from academia, uh, You're kind of prided on being very, very specific in staying with that for a long time.
And that wasn't really conducive to, um, like the way I learn and what I wanted to experience. I wanted to try a lot of new roles and for tech constantly changing to a different environment. You need to learn a new set of skills. I would say we have shifted away from. Um, the folks that, cause I would say around 2018, a lot of folks were able to get jobs as data scientists without, without a bootcamp, without a degree in data science and with only experience in the Jupyter Notebook.
But now we're in this phase of MLOps and even, uh, large language models where. You really have to think about infrastructure and think about scalability and be able to handle dirty data. Um, so we've, we've seen that, I'm sure you've seen that like with, uh, folks kind of getting new roles, going through the job interviewing process.
There's, there's just more to it than there was a few years ago, and just the names of titles are shifting and shifting. Calvin, you mentioned going from data scientist to machine learning and engineer that is the direction a lot of, uh, companies are taking, wanting their data scientists to be a bit more, um, engineering heavy and being able to handle engineering situations.
So in terms of my thought on the current job market, I. You could still go to LinkedIn. Actually, my favorite thing, um, to recommend to folks looking for new roles is go to LinkedIn, look at, look at jobs filter by the most recent jobs. If you have something like remote that's a requirement or you wanna work in a certain city, put all those options on and you could see like the latest posted ones going from company to company.
If you just think, oh, I'm interested in this company, and looking at those job rules, There's a chance that those are outdated, um, or they're not a priority. So going and seeing the latest posted jobs, like if they posted that three hours ago, I bet they're still hiring for it. And that's, that's what I tend to find.
Um, and in terms of the job market, what we're still seeing, still a lot of data science roles, but the. What's the word I'm looking for? Like the uptake. Um, like the, just like the acceleration increase for data science roles is slowing down. Uh, machine learning engineers, software engineers, data engineers, those are still really high.
But now we're seeing things like big data, MLOps, DevOps, infrastructure, um, so just. Keeping that in mind, um, you can make your search pretty broad, um, for areas of tech and then you can see the different job descriptions. Cuz once you start looking at 10, 20 job descriptions in roles you would like, you might see that there are slight differences to the job market a few years ago.
Um, and so that's, that's really where it's, it's changing a lot. Um, Chip’s session, they go into, like the most current trends. For, um, MLOps and what ML engineers are expected to know. So definitely check out that panel. Um, okay, let me see here. All right, let me go back to comments. Just navigating a few screens here.
Okay. Um, also thanks Maggie for putting in that there's a great networking event tonight. Um, In the Bay Area, uh, just this entire Arize event is a huge networking opportunity. So checker our slack out joiner, Arize slack, join our community, see who's putting in their introductions and their introductions channels because there's so many people there that.
Would make great coworkers, great managers, um, and just say hi. Uh, so our introductions channel is really blowing up right now from everyone that's participating in observe. So there's obviously a lot of speakers there, but a lot of attendees and companies who are hiring. So, Going into introduction channels and slacks is also a great way just to say hi, meet someone, and if you, you notice a lot of things in common, just comment on that.
Like, oh, same background, same university. Like notice those commonalities. I. Um, so, uh, Maggie's putting in the links for the, some of the networking sessions within the platform, uh, the Slack. And just so you know, like we do have, uh, courses as well, and you can connect with people who have taken those courses.
So just another possible networking opportunity. We have two courses, the introductory and the advanced course. All right, back to some questions here. Let me adjust my camera, make sure I'm. I'm in the middle. I have a tendency to just kind of lay back. Um, so I wanna make sure everyone can see me. All right.
All right. Next question we have about reaching out to people on LinkedIn. We're doing informal interviews, uh, oh wait, informational interviews. Um, so interesting thing about like reaching out to people on LinkedIn interviews. So I showed a few options and I'll, I'll re-share my space. Screen still sharing.
Maggie, can you see my screen? Oh, perfect. Okay. I see how it works. Yeah. So, um, a few like options for what to send to folks on LinkedIn. Um, great way to connect. And then one of the last things I wanna talk about is everything's an interview. So we, we pretty much covered. How many applicant applications to send in continuously networking with folks.
And then the job market is constantly changing. Um, but the amount of open positions, time to fill those positions is, uh, it's increasing year over year. Just the amount of roles in tech are increasing year over year. And the, the titles are changing slightly, mostly because the. Skillsets required are changing slightly.
Um, and while we are seeing, you know, large tech layoffs, um, it's a, I've talked to a lot of, uh, hiring managers and folks in the industry, and what we see is it's, it's kind of a correction to an. A bit of a over hiring in the market several years ago. Um, but those companies are still hiring, they're still hiring backend, they're still hiring project managers.
There's a lot of times still hiring entry and senior. Um, not as much in the mid-level, just for what I've been noticing in the market and what I've been talking to people about. Um, but these kind of, uh, you know, informal, um, interviews, so everything is an interview and when it might, it may sound kind of scary when I say everything's an in interview, but it is, um, if you meet someone on a networking event, um, and they're hiring and you mention you're looking.
Their thoughts are automatically, can I hire this person? Would this person make a good coworker? And it's not to scare you, scare you off, because the number one thing interviewers, like when you start a technical interviewer, the number one thing they're looking for is also, can I hire this person? No.
Like the interviewer is always on your side, so you know, they, they wanna help you. Um, and so just having, you know, just reaching out on LinkedIn, asking for 10 minutes of their time, like. You know, maybe they have that career projection that's the exact same career projection that you are hoping to get into.
Maybe it's your dream role at a company and you just want to kind of pick their brain, um, to go through their experience. Cause I guarantee you they have advice and if they do have the time, they're probably willing to hop on a, like a 15 minute zoom call just to kind of meet with you. Um, go through a few things and if they know they're hiring, it's.
Beneficial to them because if it's a great call in those 15 minutes, they could be getting. A new person on their team that will help them with their workload. Um, they might even just introduce your resume internally. Maybe they get a referral bonus and you know, if, if they think that maybe you're not the right fit for the role, they st they've still made a new connection.
Um, you know, they built out their network. Um, You know, so it's, it's normally a win-win when you have these short conversations. And the worst thing they're gonna say is no, but they'll probably still connect with you. You'll still be able to kind of follow along, um, and see if there's like any new material they're putting out.
Okay. Let's go back to some questions here. Um, next question. The pace at which they understand evolving is so fast and the job of an M l E or data scientist is changing really quickly. Uh, how do I know what skills to focus on? That's a great question. Um, those will be talked about in a chips panel in the career session, but I think for like a lot of the skills, first, it's figuring out.
Like what role you really want. So I recommend going through a lot of different job applications. Like maybe you do want, um, a data scientist role. Maybe you do want a machine learning engineering role or a data engineering role. Um, or maybe you're, you do better as a technical project manager or a product manager, um, or, uh, a similar role.
So look at those descriptions, um, see what the requirements are and think if you don't have that skill, like. Is that something you'd want to learn, wanna get involved with? Um, cuz really it's the job descriptions that are gonna tell you the most important skills to focus on. But for machine learning engineers, it's no longer enough just to.
Kind of create a linear regression model or, you know, uh, retrain, um, a neural network. You have to think about how would I get the data? Um, so starting with the data, how would I train this model? How might I continuously retrain this model? How am I gonna deploy it? And then how am I gonna monitor it in production?
Cuz it's not a lot. Enough just to build the model. Now you really have to see how it's doing in production. Um, and being able to say, you know, my model is improving business outcomes by X amount. Um, and putting business metrics, uh, and KPIs with your model as well. Okay. We got some new comments here. Um, Let's see.
Um, so Dad asked about prompt engineering. Do you see this as a bigger role in the future? I'm actually gonna link my post, um, to, uh, like the, I think it's the three main steps to prompt engineering. I do think it's becoming larger now. There's, you know, courses on how to become a prompt engineer, but it's really something that, you know, we're just gonna have to wait and see.
Um, but. There's, there's already been a few rules, which is more than there were last year. So I think it, it is picking up, but, you know, we'll see. I think it could go either way of being like the biggest role or kind of fading out or kind of merging into machine learning engineering roles. Um, Garrett asks, so before time is up, if I can expand on everything as an interview.
Um, And then he is sharing a experience here, uh, that he hopes relates to other people. I've had a lot of people who could recommend me, but none of them are hiring that, that is very common. Um, how can I get them to speak with the hiring managers and teams before I submit an application? That's a great question.
Um, and it's a great way to also have them keep you in mind as soon as they start the hiring process, cuz sometimes teams aren't hiring and then, you know, it's three weeks later and they have two roles that they're hiring for. So keeping that resume. And that's why a lot of times you might see a role, um, on a company's site that you apply for and then they say like, oh, we're not hiring for it right now.
They're collecting those resumes cuz they're about to start hiring for that. Um, and what you can do, Garrett, Is just reach out on, on LinkedIn, send a short message. Um, you know, I would say tailor these message more towards people on LinkedIn than you do when you tailor your cover letters. Um, because if, when I get a message that, you know, mentions.
I read your blog post, or I, I love, like the work you're doing or, I saw this talk when someone, uh, is making the F effort and I see like, like, oh, they actually, um, you know, know what I do. They know who I am, um, they're interested in what I had to say. Um, yeah, like I'll get on a, a 10 or 15 minute call and I do this all the time, getting on these 10 or 15 minute calls just to, to speak with them because.
That is, that is an interview. And, um, you, you can ask like, Hey, like I, I saw this job description. Um, you know, I have this experience. Do you have any thoughts on that? They may or may not, but, um, a lot of times if the conversation goes really well, if they could see you at that position, they'll offer to send and introduce the resume, um, internally.
And we know internal, uh, introductions have. I think it's about, uh, 10 to 30%, uh, return rate on getting at least that first call scheduled compared to like a one to 3% return rate on a cold application. So an order of magnitude better, um, in terms of getting, getting through that first round. And also the more connections you build, the easier it is for folks to find you on LinkedIn, like hiring managers and recruiters.
Um, and so you'll be getting more of those types of messages. Alright, everyone, I think we're out of time. Um, but I will be answering more questions in the community chat. Um, so I'm, I'm there all the time. Um, and yeah, thanks everyone. I hope you have a great day.