> It honestly gives me a huge sense of accomplishment when I get to see a customer making use of our products, and helping them investigate why something may not be working.
> I've learned that money isn't everything.
Put this at the top of your resume, send it around to a few headhunters and see what comes back.
Many (most?) developers are not very interested in supporting someone else's work. The customer seems to get it even worse. AI has turned the ego trips up to 11. You have an advantage if you genuinely feel this way about these topics and can sell that feeling to others.
> I am just much more comfortable supporting existing software than building new software.
Having technical skills is required, but you'd be doing little to no coding or design work.
> Being an interface between business and tech.
A sales engineer matches this almost by definition.
> Just working with people in general!
The job is essentially entirely about working with people, both with the customers you're trying to sell to, and with the developers who are actually making the product.
This is generally true across all roles, but first impact will be working that is more trivial for an agent or n8n flow.
So consider this and pick roles wisely, where you can see this existing for a bit. All the best.
Even if we take purely engineering, support and operations functions, there is still plenty of “human glue” in the organizations that allow for them to function - the larger the org, the more often you need to interface people between each other, and corner cases become more complex.
Big Tech which is the most vocal about their technological savviness is a great example: pretty much nothing has changed on business and cross-functional side since the 90th and earlier, all key decisions are made via Powerpoint, memos and personal contacts. Jeff’s innovation of preparing and reading a one page of problem statement and proposal at the beginning of a meeting was considered breakthrough in this field. No LLM in sight that can magically fix this [semi-ironically] beautiful mess.
And don’t get me started on B2B Sales and Partnerships.
> I get a huge sense of accomplishment seeing customers user our product’
> Interface between business and tech
You might like trying your hand in Professional Services either at a System Integrator like Deloitte or first-party in a product company (now often re-labelled as ‘forward deployed engineers’).
Their main problem space is getting customers to successfully use the product through marrying your understanding of the tech’s limitations and sweet spots with the customer’s organizational goals and culture, so daily work revolves less around deep technical design and more about org alignment and strategy (which is why SIs sometimes get a bad rap on a tech forum like HNews).
It probably pays less, and it can be more stressful when you’re billing for (and justifying) every hour, but in my experience the way it tends to be more dynamic (new projects+teams+customers often) and how you get to see very direct impact is quite rewarding.
Many products require technical product managers.
Being people oriented and a "mediocre" technical engineer is not really a hinderance and can make your focus better aligned customer needs in a high level, not getting nerd-sniped on things that don't matter.
I suspect that engineers would prefer working under someone like you who doesn't have a god-like ego, and is likely to resolve conflicts between people.
At my previous employer, the head QA engineer was the one pushing back at the PM when they were too ambitious. They basically owned the current "Is the next release still a buggy mess?" state.
Software sales engineering could be another, which probably also pays more. But, alas, more competitive.
AI is also very good at scanning legacy codebases, and reverse engineering design docs, system interfaces, and trying to reverse engineer what requirements and design tradeoffs lead to that system. It's not perfect, but already at medium software engineer level.
2026 has made him way more useful to us engineers. Before he used to send us tickets. Sometimes we delivered what the customer really wanted because of those tickets, and sometimes we needed a few iterations.
Now he uses AI to deliver pull requests that include full repro environments. These are slop, of course. Sometimes we can approve them almost unchanged. But more often they need to be completely rewritten. But even the latter is super valuable as a much more concrete representation of the bug or customer request than a ticket was.