Skip to content
Search jobs inside your AI agent. Claude, ChatGPT, Claude Code, Codex, Gemini, Cursor and Grok
Connect

RiskMaat

NewBe an early applicant

Research Engineer – Algorithms and Validation

Posted

Description

Munich · On-site · Fixed-term (12 months) · Start January 2027

Please note: We are unable to offer visa sponsorship or relocation support. Applicants must already hold the right to work in Germany.


Prove how disruptions spread, and how they can be stopped.

RiskMaat is building a Reasoning Machine for systemic risk: software that understands how organizations, critical services and their dependencies connect, and how failures can propagate through these systems.

We are looking for a Research Engineer to develop the algorithms that reason over our world model, and to show rigorously how far they can be trusted. Your work will turn a model of dependencies into answers about how disruptions spread, which measures change the outcome, and where the answers hold.


What the work involves

  • Develop scenario and reasoning algorithms that use the world model to trace how a disruption spreads to critical services.
  • Develop methods that compare measures in what-if scenarios.
  • Build reference datasets and test setups, including gold-standard sets.
  • Plan experiments, choose metrics and run statistical tests.
  • Show where the algorithms are reliable and where their limits lie.


What the role needs

  • A master's degree or PhD in computer science, mathematics, statistics or a related field.
  • At least one year of applied research or engineering on algorithms and their evaluation.
  • Excellent Python: clean, typed, tested code that others can build on.
  • Graph algorithms: traversal, paths, propagation.
  • Causal inference or scenario analysis.
  • Experimental design, metrics and statistical tests.
  • Fluent English.
  • Existing right to work in Germany.


Nice to have (suggested)

  • Network science, cascading-failure or contagion models.
  • Probabilistic modelling or simulation, such as Monte Carlo or agent-based methods.
  • Experience building benchmarks or annotation guidelines for gold-standard data.
  • Domain knowledge in cybersecurity, critical infrastructure, finance or operational resilience.


What RiskMaat offers

  • Rigorous research with real stakes. Develop algorithms whose answers support consequential decisions about critical services, and define how their quality is measured.
  • Ownership from day one. Take responsibility for a core part of the system from your first week. In a small team, your methods and evaluation choices will directly shape the product.
  • Research-driven environment. RiskMaat's project outline, EOS – Cyber Risk Intelligence, has successfully passed evaluation and been selected for further qualification under StartUpSecure (Q2/2026), a programme of Germany's Federal Ministry of Research, Technology and Space (BMFTR).
  • Transparent compensation. Fixed remuneration corresponding to TVöD Bund E 13, Stufe 2 (full-time).
  • Munich-based team. Work with a small team building the foundations of a new approach to systemic risk intelligence.


About RiskMaat

RiskMaat is building the Reasoning Machine for systemic risk.

Our software connects a model of an enterprise with models of the sectors and environment around it. It identifies critical dependencies, analyzes how failures can propagate, and reasons through uncertainty using what-if scenarios and interventions.

The objective is straightforward: understand how disruptions reach critical services, identify where they can be stopped, and direct resources towards measures that change the outcome.


Employment conditions

  • Location: Munich, Germany (on-site).
  • Contract: Fixed-term, 12 months. An extension may be possible but is not guaranteed.
  • Start date: January 2027.
  • Compensation: Corresponding to TVöD Bund E 13, Stufe 2 (currently €5,279 gross/month, full-time). Not negotiable.
  • Working language: English.
  • We are unable to offer visa sponsorship or relocation support.


Skills

Apply

See also