Neurosymbolic AI Safety & Reasoning Architect

microTECH Global LTD; Tampere; On-site; EUR 90,000 - 130,000; Full time

Job summary

microTECH Global LTD seeks a passionate researcher to advance AI safety and scalable production-ready AI systems at the intersection of AI and safety. You will tackle challenging problems and publish impactful results.

The role emphasizes designing novel architectures for multimodal data, strong emphasis on interpretable AI, alignment, and robustness, with collaboration across multicultural teams.

Qualifications

• Ph.D. in Computer Science, Deep Learning, Machine Learning, Mathematics or related field

Responsibilities

• Design and implement AI architectures for detection systems handling multimodal data (images, videos, text).
• Explore visual question answering, visual reasoning and multimodal grounding techniques.
• Develop interpretable, unbiased and aligned AI models for real-world scenarios.
• Implement state-of-the-art alignment methods (RLHF, RLAIF, Constitutional AI) for LLMs and multi-step reasoning.
• Reverse-engineer neural networks to understand internal workings and apply explainability techniques.
• Develop data attribution methods to quantify influence of data on predictions.
• Build benchmarks and datasets to measure reasoning quality and reliability.
• Create defenses against jailbreaking, prompt injections and adversarial exploits targeting planning and tool-use.
• Communicate insights to stakeholders and translate research into scalable code.

Job description

microTECH Global LTD seeks a passionate researcher to advance AI safety and scalable production-ready AI systems at the intersection of AI and safety. You will tackle challenging problems and publish impactful results.

The role emphasizes designing novel architectures for multimodal data, strong emphasis on interpretable AI, alignment, and robustness, with collaboration across multicultural teams.