Piloting renewable energies
We are QUADRA energy — and we'd like to meet you.
We bring renewable energy to the market, and we are building the company around a simple ambition: humans decide, autonomous agents do the heavy lifting. That only works with people who can think a business problem all the way through — which is why we are currently looking for a Business Analyst and a Business Architect: two humans for a very human team. This challenge is part of that search — and we'd love to hear from you.
The challenge — and the mindset we're looking for
A fictional renewables portfolio manager has published its monthly Deckungsbeitrag report — and the numbers are wrong. Exactly three deviations are hidden in the data. Your job: pull the raw data behind the report through our API, reverse-engineer how the aggregation actually works, find the three deviations, and write a spec precise enough that an autonomous agent could rebuild the report correctly.
Notice what we are not asking for: we don't need you to write the code. We expect you to bring an agent — Cursor, Claude, ChatGPT, whatever you work with — and direct it. That is the AI-first mindset we are looking for: you stay curious about the business problem, you ask the data sharp questions, and you let the machine do the typing while you do the thinking. If your first instinct when you see a broken report is “I bet I can figure out why”, you are exactly who we want to talk to.
No CV, no cover letter, no trick questions. Just a puzzle that feels like the real job — because it is. Take a coffee, take a weekend, have fun with it.
How it works
1 · Register
Enter an e-mail and a display name. You instantly get a personal access code and an API key — shown once, stored only as hashes. No verification e-mail, no newsletter — we only write to you personally if we shortlist you.
2 · Explore the API
Nine JSON endpoints with the raw material behind the report: feed-in actuals, market prices, schedules, contracts, EEG premiums, cost allocations — and the report itself.
3 · Find the deviations
Reverse-engineer how the report is built, then find where it goes wrong. Exactly three deviations exist. For each: the source, the EUR impact, the root cause.
4 · Write the spec & submit
Intent, Trigger, Data inputs, Business logic, Guardrails, Output, Success criteria. The spec is the deliverable — precise enough that an agent could rebuild the report correctly.
Ground rules
The API is built for autonomous agents, and bulk data pulls are expected — just pace your agent. These thresholds apply per participant:
- 120 requests per minute per API key, 240 per minute
per IP address. Every API response carries
X-RateLimit-Remainingso your agent can pace itself. - Bursts are capped at about 20 requests per second per IP — fire requests sequentially or with a small concurrency limit (10 parallel connections is safe), not hundreds at once.
- Above a threshold you get HTTP 429 with a
Retry-Afterheader. Wait that many seconds and continue — an agent that keeps hammering through 429s only wastes its own time budget.
Pulling the full January dataset at the allowed rate takes only a few minutes — there is no need to flood. And one promise from our side: every submission is read by a human, every decision is made by a human. The machines help us triage; they never decide.