“What’s the hardest part of being an engineer?” I braced myself for an answer I’d surely have to look up later. But Justin Yao, Senior Software Engineer, said, “It’s not the code. The code, more or less, is actually the easy part.”
I’m not an engineer. I work in Marketing. I wasn’t sure what to expect from this interview. But I sat down with Justin and Hannah Tan, another Senior Software Engineer, to talk about what growth has been like at Loop from their perspective.
Over the next hour, the two of them kept circling the same idea: technology is just the first layer of what makes solving the problem Loop focuses on challenging. What’s on top of that is a logistics industry dense with domain complexity, and a product no one has built before.
Knowing the right thing to build when there's no precedent takes judgment, and the engineers at Loop have a name for the specific kind they rely on: product intuition. Most of the time that phrase is used in relation to a product engineers use themselves. Loop's engineers rarely have that luxury. They're building for freight operators, audit teams, and a physical world of carriers and warehouses they don't personally live in, with real constraints they can't simply feel their way toward.
So the intuition they develop is a sense for what someone in a world you'll never inhabit actually needs, built secondhand by paying close attention to the people who have deep domain understanding. It isn't something you can be onboarded into or pull from a shared doc. You earn it deliberately, by building the wrong thing, watching what happens, and letting that teach you. The care you infuse early is exactly what lets you move fast later.
Everything Justin and Hannah told me over the next hour came back to that one skill: where product intuition comes from, what it looks like up close, and why it grows more valuable as AI reshapes everything around it.
The code is the easy part
On its own, a technical problem has patterns, constraints, and usually a right answer waiting at the end. A good engineer knows how to chip away at it. "If it was just purely technical, that would be easy," Justin told me. The trouble is that at Loop, technical problems are surrounded by challenges unique to the logistics industry. "It's a combination. Multiple layers of complexity."
Logistics is an old industry, plagued by fragmented siloed data that’s fed into legacy processes and tools. Every company that ships things does so a little differently, with its own documents, its own rules, and its own exceptions that’s written down on a piece of paper, if at all.
Take something as ordinary as an address for example. The same warehouse address can be written a dozen different ways across a stack of documents—abbreviated in one, misspelled in another, spelled out in full in a third. A human knows at a glance they are all the same place, software has no idea, until someone teaches it to.
That nuance is just one example in an ocean of other similar examples unique to this industry. It’s a great problem for AI to solve, and now is the time to be solving it. But as an early engineer, it’s a steep learning curve. These are physical, real-world constraints, the way a warehouse actually operates, the way a carrier actually files, and you can't guess your way to them.
That certainly surprised Hannah when she first started as a SWE I. Coming from finance, she told me the hardest part of her first months was simply learning an entirely new vocabulary of logistics terms. I remember that myself. Everyone who starts at Loop likens the early days to drinking from a fire hose. But it’s really exciting and we all are eager to dig in.
As an engineer at Loop, you are never solving a clean problem. You are solving a tricky problem stacked on a messy one, and the messy one keeps shifting underneath you while you work. On top of the fact that Loop is creating a new category. Solving a problem no one has wanted to tackle, with a platform that no one has seen before.
Building what doesn't exist yet
At most companies, you build with reference. There's a competitor's version to study, an established pattern to follow, or a quick way to tell whether you're giving customers something they recognize. But at Loop we're working from a blueprint sketched in pencil.
I know firsthand what it means to take a category creation company to market. I live the inherent challenges that come with that, but I had no idea how an engineer meets it from theirs, “How do you build a product someone wants when they can't even picture what it would look like?”
"It's very, very difficult," Justin said. You make your most informed guess about how people will use the thing, you build it, and a good share of the time, it doesn't go the way you imagined. The skill, he explained, is to stay open minded, curious, and agile. "We thought it was this way. The customers come back, they really want it this way. So we sit back and think it through again."
Hannah described the same challenge as a matter of temperament more than technique, and she learned it during a project she took on early in her tenure at Loop. She was trying to figure out how to reduce the volume of orphaned shipments for brokerage clients, since it was becoming a pain point.
Because there were two independent entry points to create the shipment, when two different artifacts were ingressed separately, it created a potential race condition when two different shipments were created to represent the same load.
Loop heard the feedback and then built a solution to centralize the creation path to eliminate the race condition and source of confusion for clients. “You learn to be comfortable with a lot of ambiguity, to keep moving on a problem when no one can tell you the right answer yet,” she explained.
This is a particularly important skill in logistics. Things change daily. We are building products with this in mind, but the people building the product need to employ the same mentality. And the engineers who don't take the correction personally, develop a unique skill that Hannah taught me. And it serves them well beyond their time at Loop: product intuition.
The key skill is product intuition
Hannah, nor Justin were familiar with Loop’s product space, or how freight audits worked, or much of what Loop did in general (it was the same for me actually). And because it's not a product any of them used in their own lives, there was no shortcut to understanding it. The only way in was to listen. So in the early months, Hannah mostly did exactly that.
She sat with how the designers and implementation teams talked to clients, and a pattern surfaced: the thing a client cared about was rarely the thing she had built the feature to do. The spec said one thing. The person using it wanted another. Product intuition is what fills that gap, and you only develop it by experiencing that over and over.
For example, Hannah noticed that clients would sometimes avoid a new feature. Not because it was broken but because it asked them to give up a process they already trusted. So instead of re-engineering the feature to be more “capable,”Hannah focused on making it less threatening. Add a tooltip that explains what someone is looking at, or a step that felt like the old way of doing things. Small, almost invisible choices that decide whether a feature gets used or ignored.
The same instinct showed up while the team was building DUX™ 2.0, a family of custom-built models that pulls clean data out of any messy logistics document. As we all know, models aren’t perfect, they are sometimes wrong, so Hannah built that assumption into the workflow from the start. A way to catch it, override it, and roll back fast. A junior engineer builds for the version where everything works. A senior one designs the path for when it doesn't.
That kind of judgement and intuition can't be handed to you on your first day. You earn it slowly, by getting things wrong and paying close attention to why, then carrying that lesson into the next thing you build. And in the age of AI, a skill like that is incredibly valuable.
AI can’t replace judgement
A lot of new graduates are uneasy right now, worried they spent four years learning something a machine can do. It's a fair worry. So I asked Justin and Hannah how they see AI changing their field, and they didn't flinch, "We just don't know," Justin said.
Justin watched the tools go from simple autocomplete to something seriously capable in a short span of time. What that means going forward is a question mark. Maybe engineers move up a layer and spend more time directing the work than writing every line of it. Maybe it’s something else entirely.
AI is extraordinary at what has already been done. It can complete the known, the documented, the pattern it has seen a thousand times. What it cannot do is decide what should be built in a domain this messy, or sense how a particular customer will use the thing once it exists, especially when that customer lives in a physical world the model has never touched. "AI will probably have a lot of knowledge on the logistics domain," Justin said. "But it won't know how your customers actually use your product. That's where you come in." That gap, between what has been done and what needs to be done, is exactly the gap product intuition fills.
Working at Loop has never once let Justin or Hannah coast. The carriers change their rules. The customers change their minds. Now the models change week to week too. Having to adapt faster than feels comfortable is the exact thing three years at Loop has already trained into them. Thinking more broadly, perhaps this is the underdeveloped skill that engineers need more than ever in this day and age.
What they walk away with
I came in thinking engineering was about writing code. An hour with Justin and Hannah took that apart. What I left with was something much more rich and thought provoking.
The growth was real. Both of them are senior engineers now, trusted with work that would have unsettled them three years ago. But the promotions were the residue of something harder to put on a resume: the judgment to build the right thing in a domain that won't sit still, for products that don't exist yet, without an answer key to tell you when you've got it right.
That judgment is what Justin and Hannah were really describing the whole time. It's what every part of the job trains into you, and no tool can replace.
AI will keep getting better at the work that's already been done. The work that's left, is the work that makes an engineer, and it compounds its value.
The engineers who thrive at Loop are not the ones who want the answer waiting in the back of the book. They're the ones who feel a small pull toward the blank space where the answer should be. If that's you, you'd be in good company here.
Check out our career page and apply today.
