University of Delaware · Electrical & Computer Engineering

Getting Started and Navigating a Non-Linear Career in the Age of AI

You just need the next step.

Lauren Saxton /Sr. AI Engineering Manager, Veeam /AI Advisor /UD Class of 2013

Before anything encouraging

The market you're walking into is hard.

Entry-level postings are down.
More applicants per opening than when I graduated.
Companies say out loud that AI now covers a lot of the work junior engineers used to be hired to do.

If you're anxious about that, you're not being dramatic. You're reading the situation correctly.

So who am I to say there's room for optimism?

Right now

Sr. AI Engineering Manager — Veeam, ~8,000 people

AI advisor to startups

Standing here

Also has been

Consultant · Frontend engineer · Backend engineer

Graduate student · Cofounder · Product manager

Laid off

On the thing a lot of you are afraid of

Getting laid off is not the end of a career. Sometimes it's the most interesting turn in one.

Once

I came back at nearly double the salary.

Once

I spent a month living in London and traveling Europe first.

The Elizabeth Tower clock face and spire, with Westminster Abbey behind it — Big Ben

On horseback in a cobbled London mews courtyard — Hyde Park riding

Two people at the base of the Eiffel Tower, looking up its full height — Eiffel Tower

Thirteen years

I've already lived a lifetime and grown in ways I couldn't have predicted.

My career has let me travel the world and build a life that works for me.

I want you to leave here believing that's available to you too.

Standing in an orange harness suit on a walkway near the top of a tower, high above a harbour city — Auckland Sky Tower

Sitting on the crest of a tall red sand dune, with more dunes running to the horizon — Sossusvlei dunes

A thing I actually believe

The five year plan is dead.

Five years is long enough to earn another full degree. In a world of self-driving cars and AI coworkers, who's confident enough to say what anything looks like in five years?

What replaces the plan

With the right perspective and a few small deliberate actions, you can absolutely control your choices — and build a career that keeps you engaged through every phase of your life.

Where we're going

My path. Not a straight line, and not planned.
What's actually changed in hiring now that AI is in the mix — from someone who does the hiring.
What to do about it. Practical things you can start immediately.
Questions. Plenty of time.

Part one

01

My path

You won't be the same person in five years that you are right now. So why should your career be?

Careers I'd have if I'd trusted my earlier selves

Childhood passion

Racehorse jockey

Early college ambition

Fire protection engineering

Firm belief at 22

“I'd hate working remotely”

Now I work wherever I want, and I've lived in different parts of the country and the world.

Stop 01

2013

I had no idea what to do next.

I wasn't a great student. I wasn't good at electrical engineering and it never came easy to me. But I loved my computer science classes. That was the first real spark. I knew a lot more about what I didn't want than what I did.

Wanted: Money · To move away from Delaware

Filled pads are stops. Ringed nodes are vias — the turns that only happened because of a person.

Stop 02

2013

Avanade — New York City

Consultant · rotating clients and projects

Two reasons. One: they were willing to hire me, unlike the flashier names I applied to. Two: they'd pay to relocate me. Software is a huge world — picking what kind of engineer to be is like picking a major all over again. Consulting bought me time and exposure while getting paid.

Then what I wanted changed: Stability · Depth

A very small studio apartment: a loft ladder, a sofa, an armchair by the window, and a refrigerator crowding the foreground — The shoebox

Consulting is not always glamorous global travel. Sometimes it's New Jersey four days a week for years.

Stop 03

2016

Match.com — Texas

Frontend engineer → Senior · three years

The first time I doubled my salary. I want to be clear about how it happened: I doubled it by telling someone I was interested. That's it. That was the whole strategy.

Then what I wanted changed: I was bored

There is nothing less like New York than Texas.

Stop 04 · a via

2019

Derek — over lunch

One Technologies

Backend (I'd only done frontend) · finished my master's · built the ML team from scratch

Two years building a team and an architecture from nothing. That was the foundational moment of my career — and I would not have been in the room for it if I hadn't had lunch with a friend.

Remember his name. I'll come back to him.

Stop 05

2021

NerdWallet — pre-IPO, then public

Engineering manager · blackout periods · reorgs

Going public was electric. It was also the first time I managed a team through a lot of anxious questions I couldn't fully answer. Then my whole team was laid off when the company changed direction on ML.

That's tech: You learn to hold the highs and the lows with the same hands

A volcanic crater glowing orange at dusk, with a steam plume rising into purple clouds — Volcanoes National Park

Palm trees and dense jungle framing a view of surf breaking on a rocky coast — Big Island

Stop 06 · a via

2022

A former colleague — a phone call

Cofounder

One year · no guaranteed paycheck · genuinely scary

I did it anyway, because I wanted to know what it felt like to be in the driver's seat. I learned an enormous amount. I also learned I wasn't passionate enough about the industry we were in. So I ran the experiment, got my answer, and moved on. No drama.

Stop 07

2023

Pachyderm → acquired by HPE

Technical product management · a deliberate two-year window

Something genuinely different. I gave myself two years to find out if product was for me. At the end of two years I had my answer: I missed engineering management. So I went back to it.

Stop 08 · now

2026

Veeam

Sr. AI Engineering Manager · advising startups · standing here

Eight stops. Three of them happened because of a person. None of them were on a plan.

A stone arched bridge spanning tall sandstone rock pillars above a forested valley — Bastei

A river embankment with moored boats, and a castle and cathedral on the hill beyond — Prague

The woman who graduated from this school with an EE degree she didn't love would not have understood a single word of my current job title.

That's the whole point.

Part two

02

What's actually changed

Let's separate the realities from the ghost stories.

The fear is not unfounded. But —

Changing is not the same as closing. The difference between those two things is where your whole career lives.

You wouldn't try to survive in a desert using techniques you learned in a rainforest. Step one is learning the environment you're actually trying to succeed in.

The honest bad news

−19%

Widening — it was −15% a year ago

Employment for workers aged 22–25 in the occupations most exposed to AI, relative to where it would be if it had kept pace with older workers in those same jobs.

Stanford Digital Economy Lab · ADP payroll data covering millions of US workers · update released August 2026

Two things the headline leaves out

It is not happening through layoffs

It's happening through reduced hiring. Companies aren't firing junior people. They're not opening the roles in the first place.

That's a real problem, and it's the one you're feeling.

It is not happening everywhere

The declines are concentrated in jobs where AI substitutes for what a person was doing. In jobs where AI complements the person instead, employment is flat or growing.

The most actionable fact in this entire talk

The question is not whether you work near AI. The question is whether the work you do is the kind AI replaces or the kind AI makes more valuable.

The frame worth memorizing

Codified knowledge

What's written down and repeatable

What junior people were hired for

Exactly what these systems reproduce well

Tacit knowledge

Judgment, built by doing it and getting it wrong

What experienced people are hired for

What the machines can't do yet

Your job is to build tacit knowledge faster than the old timeline assumed you would.

The half that doesn't make headlines

Software job postings did fall off a cliff. Then they started climbing.

Up about 15% — while job postings overall have gone down. The market is recovering.

Indexed to pre-pandemic = 100. Anchor points are the reported figures; the line between them is illustrative.

The catch

The building is filling back up. The front door is still narrow.

71%

of the recovery is senior roles

37%

is jobs with AI in the title

−25%

still below pre-pandemic postings

Both things are true.

Two numbers that never get quoted next to the scary ones

Early-career pay

#1

of 73 majors the NY Fed tracks. Computer engineering, around $90,000. Computer science right behind it.

Underemployment

Working a job that doesn't require your degree.

Federal Reserve Bank of New York — labor market outcomes by college major

The degree didn't stop working. The front door got harder to open while the building kept growing.

Those are different problems. The second one is solvable.

The part I most want to say to this room

Everything happening in AI runs on hardware that somebody has to design, and on power that somebody has to deliver.

Hardware · Semiconductor Industry Association

The US chip workforce is growing by 115,000 jobs by 2030.

Not enough people who understand electronics, RF, embedded systems. About a third of that gap is engineers with four-year degrees.

An industry that expects to be tens of thousands of engineers short — in precisely the disciplines taught at this college.

Power · Gartner

Data center power demand

40%

of AI data centers will be constrained by power shortages by 2027.

The detail I find most telling

The entire AI buildout is now bottlenecked on the electrical grid.

That is an electrical engineering problem. It is enormous, it is urgent, and there are not enough people to solve it.

I'd rather say this than have it sit unsaid in the room

Not everyone thinks this buildout is a good thing. Right here in Delaware, data center demand is part of why electricity prices have gone up, and there's legislation in Dover right now about who should pay for it. Some of you may have real objections — environmental, political, or both. I'm not going to tell you those objections are wrong.

One legitimate answer

If you don't believe in something, don't spend your career on it.

Also legitimate

If you think it's being done badly, that's a reason to go work on it. The people who fix a problem are usually the ones close enough to see it.

Just make it a choice instead of an accident.

When someone tells you the market is bad, the right response is —

Which market?

From where I sit — the person reading the resumes

I am not hiring anyone to write syntax.

That used to be a real category of junior work. It's the biggest change, and vague reassurance is worse than honest difficulty.

What I'm looking for instead

Judgment

Can you tell when the AI is wrong? It is confidently wrong constantly. Someone who can't catch that is a liability rather than an accelerant.
Can you move an ambiguous problem forward with whatever tools exist, without being told every step?
Can you explain why you built it this way and not another way?

That's a higher bar than the one I had to clear in 2013. I won't pretend otherwise.

The flip side of the same change

The distance between learning something and having built something has never been shorter.

Part three

03

What to do about it

I don't have magic bullets. What I have are techniques I spent a career learning, and what I see from the other side of the hiring table.

1

You don't need the whole plan. You need the first step.

Be open to opportunities you hadn't considered. Trying something is not the same as committing to it forever. You do not have to land the perfect job out of the gate. Almost nobody does. I certainly didn't.

2

Get honest about what you want right now — and what you'll trade for it.

My answer in 2013 — wanted

Money

A chance to leave Delaware

Didn't care about

A big-name company

Working remotely

That flexibility is what gave me options. Neither answer is wrong — just be real about how yours changes the math.

The one I wish someone had told me at twenty-two

Just ask.

Not a hint. Not a subtle nudge you hope somebody picks up on. Not a ten-page case you build before you'll say the thing out loud. Ask.

I doubled my salary by telling someone I was interested. That was the whole strategy.

Asking isn't a burden — most people like being asked

Ask a professor about their research.
Ask an engineer how they got their job.
Ask a company what they'd want to see from someone at your level.

The worst outcome is a no — and a no tells you where you stand. Silence tells you nothing.

3

Use the advantage you actually have.

You are, by definition, the most recently educated people in the room on a lot of this technology and the theory underneath it. For those of us already in industry, keeping up happens on top of a day job — or on top of unlearning old habits.

You have real knowledge right now, even though it doesn't feel like it, because it doesn't come with a job title attached yet.

“Learn AI” is useless advice. Here's the specific version.

Build things. Not tutorials.

Things with an actual end, that actually work, that somebody other than you could use. Use AI tools while you do it — and pay close attention to the moments where the tool confidently hands you something wrong.

That noticing is the skill.

It has never been cheaper to build something real. What would have taken a team and a year can now be done by one person in a few weeks. Raising money is a different story and harder than the headlines suggest — but you don't need funding to build something, and a thing you actually built and shipped is worth more in a conversation with me than almost anything else on a resume.

You can start collecting evidence of it before anyone hires you at all.

4

Build connections.

The more technical and automated this industry gets, the more human connection matters.

Referrals are the skip-the-line ticket in a sea of thousands of keyword-matched resumes. But you can't cold-email a hundred recruiters asking to be recommended — everyone can smell it. Referrals come from community, and community is built on genuine relationships.

Get to know your classmates. Your professors. Your parents' friends. Tell people what you're working on and what you want.

Every real turn came through a person

Derek — over lunch

Put me in the room where I built my first ML team. I didn't meet him strategically. I just liked how he managed people.

A former colleague — a call

The startup. The first thing I ever did without a guaranteed paycheck.

Two names I passed along at 22

Agilent, up the road in Wilmington. One still works there thirteen years later. The other built a management career and now leads bigger teams elsewhere.

I was twenty-two. I had no idea what I was doing. The most valuable thing I did that year was pass along two names.

You never know where the opportunities come from. You only know they come from people.

Be good to them now. Not because it pays off later, although it will — because this work is a lot better when you're not doing it alone.

Pulling it all together · the practical steps

Run experiments, not plans. Pick what you're curious about → give it a real window → ask honestly whether you're still learning → keep going, or run a different experiment.
Build tacit knowledge now. You are the most recently educated people in the room — use it. Start building while the theory is fresh. Judgment only comes from making something, being wrong, and finding out why. That's what people are hiring for.
Invest in your community before you need it. Get to know the people in your sphere now, expecting nothing back. And that sphere reaches much further than you think — it's people who know people who know people.

The market is tough. And —

You're graduating with a degree that is genuinely in demand.
Into an industry short exactly the kind of engineers this college produces.
At a moment when the tools to build something real have never been more available to you.

You don't need the whole map. You just need the next step.

Stepping off the edge of a tower platform in a harness, high above a harbour city — The first step

Grinning mid-descent alongside the tower, sunglasses on, throwing a peace sign — The rest of the way

Let's open it up

Ask me anything.

Or tell me about something you're working on.

lauren.talk

Lauren Saxton /hello@lauren.talk /lauren.talk /linkedin.com/in/lauren-saxton-37879556