Data Analyst I Courier RTO Global HQ

YOUR MISSION
Real-Time Operations keeps Glovo’s deliveries moving: when an order stalls, a courier stops, or a store goes quiet, RTO is what notices and acts.
Your mission is to find what is actually causing those failures and then build the thing that fixes them — the analysis and the automation, not one handed to someone else to build.
You will work at the intersection of operations analytics, AI agent quality and automation engineering, on problems that are genuinely unsolved.
THE JOURNEY
Own operational analyses end to end — take an ambiguous problem, find the real driver rather than the most visible symptom, and turn it into a decision about where to focus.
Build and run the automations that come out of that analysis: real-time detection, the action it triggers, the monitoring that proves it still works, and the documentation someone else can operate it from.
Work on the systems around our AI voice and chat agents across fourteen markets — automated conversation audits, the loop that turns an unanswered question into a knowledge base update, and the evaluation that says whether a prompt change helped or hurt.
Design and read the experiments behind what we ship: choose the unit of randomisation, pick the metric that reflects the real outcome, and report the result honestly, including when it says the thing does not work.
Build processes that send an input to an LLM and need a specific, reliable output back — together with the checks that confirm the output is what we expected.
Hand findings to the right owner when they sit outside RTO, with enough evidence that they can act on them.
WHAT YOU WILL BRING TO THE RIDE
Must have
Advanced SQL — window functions, CTEs, joining across large event and transactional tables, and the habit of checking what a column really contains before building on it.
A structured, sceptical approach to problem-solving. You look for the reason your own conclusion might be wrong before someone else does.
Hands-on experience building something end to end that ran in production and that other people depended on.
Good written and verbal English. Most of your stakeholders are operations people, not analysts.
Nice to have
Python (pandas or equivalent), experiment analysis, BigQuery, n8n or similar orchestration, Google Apps Script, Looker, voice or chat AI agent platforms such as ElevenLabs, and a background in last-mile delivery, marketplaces or logistics.
Practical experience using LLMs inside a process that had to produce a reliable output — prompt design, structured outputs, and a way of checking the result.

Employment type
Full-time

Work arrangement
Hybrid