Daniel Huber, AI Engineering Lead - AI Foundry / Tech Lead, Financial Technology @ Leonteq · Zürich
Taking LLM apps and agents from business idea to governed rollout in a regulated financial institution - platform, delivery, and enablement.
Focus: Agent Engineering · RAG · On-prem LLM Serving · Governance & Auditability · Observability
Stack: Python · LangGraph · LiteLLM · Langfuse · C# · 10+ yrs
# AI systems journal, topics auto-collected by agent, deployed via GitHub Actions
Leading Leonteq's AI Foundry: the internal AI platform, the LLM applications and agents running on it, and the AI strategy and team behind it. I embed with business teams, scope their ideas into agent workflows, prototype fast, and carry each build through audit review toward production go-live. Hands-on engineer with 10+ years in financial technology, alongside leading the Sophis core platform team.
AI engineering lead at Leonteq, driving the AI Foundry: the internal platform teams use to build and govern LLM applications and agents. I proposed the AI strategy to the CIO and lead the deployments running on the platform, hands-on in Python.
The job is a delivery loop: embed with business teams, scope their ideas into agent workflows, prototype fast, and carry each build through governance and audit review toward production. Alongside it I lead the Sophis core platform team and tech-lead the client-onboarding platform.
Over the past decade I've progressed from analyst to director, working across pricing/analytics, regulatory projects (MiFID II, SFTR, FRTB), and platform migrations. Now focused on bridging the gap between traditional fintech and AI-powered workflows.
Leading the AI Foundry, Leonteq's internal AI platform and service layer. Three delivery streams run on it: a knowledge app, agent workflows built from business requests, and agents embedded in existing internal applications. I lead all three, hands-on in Python; larger builds run with delivery partners under my project lead.
Tech lead for the client-onboarding platform, automating compliance and regulatory workflows to reduce manual effort and onboarding risk.
Leading a team of 4 developers across Zürich and Lisbon on the firm's core trading and risk platform. Primary interface between the technical team and business stakeholders.
Implemented generic booking model for mirror booking, reducing client onboarding complexity. Built Grafana/Kibana monitoring. Key contributor to LIBOR migration.
Progressed through multiple roles focusing on automation, risk management, and regulatory compliance.
Focused on reporting automation and distributed system implementation. Automated creation of listed instruments via SmartCo integration.
Personal projects applying ML/AI and agent development to financial technology
Independent research into how LLMs deployed in Swiss financial institutions can be audited against FINMA supervisory expectations (Guidance 08/2024). Four threads: SP-Benchmark, a 56-question assessment of LLM comprehension of Swiss structured products scored by dual LLM judges; per-decision audit trails built from attribution graphs and token-level introspection; behavioural drift detection using sparse autoencoders; and a directional-error transcript scanner that probes model conduct under adversarial pressure.
An AI systems journal where topics and news are automatically collected by an agent, pushed to the repo via GitHub Actions, and trigger a fresh deployment, no manual writing involved. The interesting part is the architecture: agent → content extraction → commit → CI/CD → live site.
A self-improving AI trading system inspired by Karpathy's autoresearch: 8 AI agents compete on real Kraken Futures crypto perpetuals while the system improves them overnight. It mutates agent configs, tests variants on past data, and keeps the winners. Agents reflect on mistakes and build rules to avoid them. When problems persist, the system writes its own code fixes. The question: will agents that learn from their mistakes actually outperform simpler ones?
A live near-Earth asteroid explorer built around two AI systems: a tool-calling chat agent that drives a 3D solar-system scene (selecting asteroids, changing views, running impact simulations), and an autonomous monitoring agent on a 4-hour cron using an advisor pattern (a low-cost Haiku executor that escalates higher-stakes calls to an Opus advisor), with persistent memory and email alerts. Data from NASA's NEO API.
Study project combining textual data mining with technical analysis and portfolio theory for automated trading strategies.
GPU-accelerated image processing using wavelet transformation and parallel algorithms for progressive graphics file format optimization.
Portfolio management and trend analysis system developed for Dufour Capital AG, combining financial mathematics with web technologies.
Agent & RAG platforms are the area I focus on most. To show how I approach these systems end to end, here's a generalized reference architecture I put together.
A vendor-neutral enterprise RAG & agent platform. Pick a request flow to trace it through the stack.
Generic reference design, shown for illustration, not a depiction of any specific production system.
Bachelor's degree in Computer Science with focus on practical software engineering, algorithms, and system design. Completed several industry-relevant projects including algorithmic trading, GPU programming, and financial portfolio management systems.
Mathematics studies at ETH Zürich. Built a foundation in linear algebra, calculus, probability theory, and mathematical analysis. Transitioned to applied computer science to focus on practical implementation of mathematical concepts.
Neural networks, CNNs, RNNs, sequence models, and deep learning optimization techniques.
View Certificate →Agent architectures, tool integration, Google ADK. Built a trading strategy agent for the Kaggle competition.
View Badge →Supervised/unsupervised learning, neural networks, deep learning fundamentals.
View Certificate →Linear algebra, calculus, probability, statistics.
View Certificate →Investment strategies, risk management, complex financial instruments.
View Certificate →Open to conversations about AI systems, fintech engineering, or agent architecture.