Machine Learning Certification Researcher

About the Role

Imagine pioneering the mathematical and logical foundations that enable AI to fly safely in real-world, safety-critical aviation. As a Machine Learning Certification Researcher, you will address one of the most fundamental challenges in modern AI: proving that ML models are safe, reliable, and mathematically sound for deployment in autonomous flight. Your work will bridge the gap between cutting-edge statistical learning theory and formal aviation certification frameworks, building upon previous work (EASA CoDANN I and II, and FAA report DOT/FAA/TC-21/48), and putting real-world systems to the test.

At Destinus, we are shaping the future of defense through high-speed, autonomous flight technologies. We develop next-generation unmanned aerial systems that deliver greater speed, precision, and efficiency, supporting governments and defense organizations in mission-critical operations worldwide. Daedalean AG, the AI and autonomy hub of the Destinus Group, plays a key role in this vision—developing safety-critical AI systems at the intersection of advanced research and real-world aviation. Enabling machines to perceive, understand, and make decisions in real time, we combine aerospace engineering and artificial intelligence to build the next generation of autonomous flight systems, driving innovation where it matters most.

What You'll Do

• AI Certification Strategy: Analyze our existing AI certification strategy and help adapt it—or pioneer entirely new strategies—to certify different supervised ML.

• EASA Framework Alignment: Connect mathematical guarantees, technical arguments, and model evaluations to the EASA W-shaped process for AI certification in aviation, ensuring our research aligns with the latest regulatory guidelines.

• Theoretical Foundation & Modeling: Conduct research into statistical learning theory, generalization bounds, and mathematical guarantees for deep neural networks operating under dynamic real-world conditions.

• Uncertainty Quantification: Apply statistics and stochastic processes to model environmental uncertainties and sensor noise, and to quantify confidence in performance metrics.

• Literature & Proof Evaluation: Critically evaluate academic literature, mathematical proofs, and engineering assumptions with peer-review rigor to identify safety limitations and theoretical gaps in state-of-the-art AI.

• Regulatory & Audit Justification: Construct robust, logical arguments and technical justifications that translate complex, high-dimensional empirical data into structured evidence for safety audits and certification frameworks.

• Algorithmic Validation: Write clean, reproducible code in Python to test theoretical hypotheses, validate certification mechanisms, and benchmark model safety frameworks.

Who You Are

• Critical Thinking & Healthy Skepticism: You possess a strong spirit of questioning everything. You naturally challenge conventional ML assumptions, probe for hidden edge cases, and refuse to take empirical accuracy at face value without rigorous proof.

• Logical Argument Building: You have an exceptional ability to construct robust, airtight logical arguments and technical justifications. You can translate complex empirical observations into structured, defensible evidence that satisfies strict

engineering and regulatory scrutiny.

• Thrives in Ambiguity: You are motivated by hard, unsolved mathematical problems at the intersection of theoretical computer science, safety engineering, and real-world flight operations

Education

• Currently pursuing or recently completed a Master’s, or PhD in Computer Science, Mathematics, Theoretical Statistics, Electrical Engineering, Physics, or a related quantitative research field.

Machine Learning Theory & Mathematics

• ML Theory & Statistical Learning: Deep understanding of statistical learning theory, generalization bounds, and the mathematical foundations of deep neural networks.

• Statistics & Stochastic Processes: Strong background in advanced probability, statistical modeling, and stochastic processes.

• Python Engineering: Proficiency in Python with the ability to write clean, reproducible code (can also be learned on the job provided strong theoretical foundations and basic Python coding).

Research Rigor

• Scientific Research Methodology: Demonstrated capability or background in conducting comprehensive literature reviews, identifying gaps in current state-ofthe-art AI, and synthesizing complex academic concepts into actionable engineering frameworks (can also be learned on the job if you possess a strong quantitative baseline).

• Peer Review Expertise: Ability or experience in the peer-review process—capable of critically evaluating mathematical proofs, algorithmic assumptions, and research papers to assess their validity and limitations.