CTO / Quant Engineer

Enzian Labs AG seeks a Senior CTO / Quant Engineer in Zürich to build an AI backbone for private equity, focusing on probabilistic modeling and production ML systems.

The company is developing an artificial intelligence platform designed to serve as the core infrastructure for the private equity sector. This role involves leading the technical strategy to transform raw, complex financial data into reliable, scalable AI-driven insights.

Responsibilities

• Manage the complete machine learning lifecycle, moving models from experimental Jupyter notebooks to robust, production-grade systems.
• Design data ingestion pipelines that can process noisy, real-world inputs such as scanned PDFs, inconsistent reports, and incomplete data.
• Customize large language models and vision models for specific financial applications.
• Implement cost-effective training techniques, including LoRA, quantization, and efficient training loops, to make large-scale AI feasible.
• Use Bayesian inference and probabilistic programming to quantify uncertainty in private market valuations.

Requirements

• A quantitative degree (MSc or PhD) in computational finance, statistics, applied mathematics, physics, or machine learning.
• Strong proficiency in probabilistic programming frameworks such as JAX, NumPyro, or PyMC.
• Demonstrated experience constructing data pipelines for messy, real-world datasets.
• Hands-on expertise in fine-tuning large models for specific domains using tools like Unsloth or HuggingFace.
• Proven track record of deploying production-ready machine learning pipelines.
• Deep familiarity with the modern AI stack, including PyTorch, HuggingFace, and Unsloth.
• Practical experience working with foundational models and successfully fine-tuning LLMs.

Nice to have

• Experience with agent-based modeling and economic simulations.
• Understanding of key financial metrics such as NAV, IRR, and fund structures.
• Prior work experience in the private equity industry.

About the company

Enzian Labs AG prioritizes rigorous statistical thinking, preferring distributions over single-point estimates. The team is committed to maximizing model performance while operating within the constraints of a startup budget.

• Emphasis on thinking in distributions rather than relying on point estimates.
• Discomfort with systems that provide numbers without credible intervals.
• Focus on achieving high model performance despite limited financial resources.