Technical Support Engineer at Glide · LumiJobs

Responsibilities

We are hiring two Technical Support Engineers to join the small team that owns production quality for the entire platform.
This is not a ticket-routing job. Our support engineers own bugs end to end: they take an ambiguous report from a financial institution, drive it to a real root cause across the frontend, backend, and a stack of banking integrations, ship the fix themselves, and prove it works.
What makes the job unusual is how it gets done. This team runs AI coding agents as a production system, not as autocomplete. A triage agent proposes a root cause. The engineer decides whether it is right, and steers it when it is not. The agent writes the code, a second agent reviews it, and the engineer reviews, tests, and records the evidence before anything ships. The team built the tooling that makes this possible: internal MCP servers, slash commands for each stage of the workflow, an auto-triage loop feeding the bug board.
The result is a very small number of people holding quality for a platform serving a lot of institutions.
So we are not hiring for typing speed. Agents are fast and often confidently wrong. We are hiring for the judgment they do not have: recognizing when a plausible answer is the wrong one, knowing what not to change, and being able to prove a fix rather than assert it.
Own bugs and production issues end to end: reproduce, diagnose, fix, verify, and confirm the fix held
Drive ambiguous reports to a true root cause across our web apps, APIs, and banking integrations
Orchestrate AI agents through the full diagnostic and repair loop, and catch them when they are wrong
Ship your own fixes as reviewed pull requests, with evidence attached to every ticket
Extend the agent tooling itself. The workflows, MCP servers, and triage agents this team runs were all built in-house, and you will build the next ones
Write the plain-language root cause explanation that goes back to the customer
Find and fix classes of bugs, not just instances. Strengthen alerts, runbooks, and guardrails so the same issue does not come back

Qualification

Strong TypeScriptYou debug by narrowingYou manage yourselfWhat the interview process looks likeWe tell you up frontIntro callA build exerciseIn person in NYCExperience at a fintechFamiliarity with Next

Required

2 to 6 years writing production software, ideally somewhere you were close to real customers and real incidents
Strong TypeScript. Comfortable moving between a React frontend, a Node backend, SQL, and a third-party API in the same afternoon
You debug by narrowing, not guessing. You can read logs, traces, and a database and form a hypothesis you then try to disprove
You already work with AI coding agents daily, and you have opinions about where they fail. You can tell when an agent is confidently wrong, and you know what you would never hand one
You can explain a technical failure to someone who has never seen the code, without condescension
You manage yourself. Nobody will assign you work hour to hour, and you will flag a slip before the deadline rather than after
You are comfortable when the process is still being written, because you will be the one writing parts of it
What the interview process looks like
We tell you up front, because ours is unusual and deliberately so.
Intro call, 30 minutes. A conversation with our CTO about the hardest bug you have ever fixed, and how you actually work with AI tools.
A build exercise. We give you a spec for a small service. Expect about four hours, with a week to return it. Use whatever AI agents you like, and send us your session logs. We read your evidence and your written decisions before we read your code.
In person in NYC, 2 hours. You defend what you built, then we introduce bugs into your own codebase and you debug them live on your laptop with your own tools, with two of our support engineers watching.

Preferred

Experience at a fintech, or anywhere regulated, where a bug has consequences beyond a bad user experience
Familiarity with Next.js, tRPC, AWS, Sentry, or a structured logging platform like Axiom or Datadog
Experience with multi-tenant systems where behavior is driven by per-customer configuration

Technical Support Engineer

Responsibilities

Qualification

Required

Preferred

Benefits

Company

Glide

More roles at Glide