Engineering Manager for Groupon's Internal Next-Gen App

Groupon IQ is the internal platform that powers deal creation, QA, and publishing across the company. The 7-person team can't keep up with the backend incidents IQ generates today, and downstream tribes feel every outage. When this role works, IQ stops being a recurring incident and becomes the boring backbone other tribes rely on.

Who you'll work with

You'll report to Pavel Jakl, B2B Tribe Director. You'll lead a 7-person engineering team across Prague and Bangalore, working in TypeScript on a shared monorepo. Pavel codes himself and expects you to. People management is a side responsibility on top of the engineering work, not a substitute for it. This won't work for anyone who needs a settled roadmap, or who treats AI tools as a threat. Both the org and the work move faster than that.

What you'll be doing

• Owning Groupon IQ's backend reliability hands-on, debugging distributed-system failures, fixing the root cause, and writing the post-mortem that prevents the next one
• Generating most of your own code through AI agents (Claude Code, Cursor, Codex, your choice) while doing the review that catches what the model doesn't see
• Setting and enforcing the code-review bar for AI-assisted output across the team, so silent timeouts and broken migrations get caught before they ship
• Driving technical direction across the shared TypeScript monorepo so a change in one app doesn't quietly break three others downstream
• Mentoring the 7 engineers reporting to you toward the ownership posture Pavel hires for: take the problem, don't wait for it to be assigned
• Partnering with the Encore team and the marketing and BI users downstream of IQ, so platform changes land predictably for everyone who depends on them

What success looks like

Day 30. You've shipped at least one backend fix to a recurring IQ incident, met all 7 engineers in 1:1s, and identified two reliability risks the team hadn't written down before you arrived.

Day 60. You've defined and rolled out the code-review standard for AI-generated changes, including the failure modes the team now actively looks for. At least one of the tech-debt items Pavel flagged on Day 1 is closed.

Day 90. IQ's incident rate is measurably down against your Day 30 baseline. At least one engineer on your team is operating with the ownership posture Pavel hires for. You've made a technical call that affected another tribe and defended it.

What this role demands

• Hands-on full-stack (preferred) or backend TypeScript depth in distributed systems. You diagnose timeouts, race conditions, and broken migrations at the source, not just patch the symptoms.
• Daily AI-assisted engineering practice. Most of your code today goes through Claude Code, Cursor, Codex, or equivalent, and you can spot the failures the model misses.
• Monorepo experience at scale (Nx, Turborepo, pnpm workspaces, or similar) where you've personally driven structural decisions about repo layout, build performance, or shared-package boundaries
• A hands-on coding role with people leadership as a secondary responsibility, already lived, not aspired to. You're the lead engineer who happens to manage a few people, not the manager who used to code.
• Ownership instinct outside the corporate norm. You've taken problems that weren't yours and finished them, because they needed finishing.
• E-commerce, marketplace, or high-traffic consumer product context where reliability translates directly to a user or business outcome you can name

How we operate

Extreme Ownership. Nobody here waits for a ticket. The engineer who finds an incident first owns it end-to-end, regardless of which tribe shipped the original code.

Impact Obsessed. Every technical call gets justified against the marketing or BI user it affects, not the elegance of the design. A rewrite that doesn't reduce incidents or unblock a downstream user isn't worth shipping.

Disciplined. AI generates most of the code on this team. Discipline is what stops that velocity from becoming a stability problem: every AI-generated change earns its review like any other, and reviewers are paid to be skeptical of the model.

The hiring process

1. Phone screen with your recruiter (30 min) — Career trajectory, role fit, comp alignment
2. Foundry Challenge — A role-specific scenario sent after your screen
3. Hiring manager interview (60 min) — Your Foundry Challenge review, in-depth technical disucussion, dive into Operating Principles fit
4. Panel interview (60 min) — Remaining Operating Principles and hands-on technical depth assessment
5. Debrief and decision

#LI-KH1  #LI-Hybrid #marci