Engineering Lead at DQC
Andreas Kanz
I help teams decide what to build with AI, design the systems around it and take them into production. My experience spans agent-based products, applied machine learning and data platforms, with hands-on work from architecture to implementation.
Expertise #
AI & machine learning #
Designing agent workflows and predictive models, evaluating behavior and integrating AI into products. Model selection, data requirements and trade-offs in privacy, cost and reliability.
Software & data systems #
Designing applications, APIs and data infrastructure. Building and modernizing production systems, including the engineering needed to turn research code into maintainable software.
Engineering leadership & delivery #
Guiding technical decisions, growing engineering organizations and remaining hands-on in delivery. Establishing the development practices, verification and feedback loops that help teams build reliable software.
My technical work ranges from agent orchestration and LLM gateways to backend architecture and data processing. I use profiling, benchmarks and runtime evidence to find bottlenecks, improve efficiency and verify that changes work.
Selected experience #
AI systems & engineering leadership #
At DQC, I lead engineering and work directly on architecture and core product features, building AI systems that help enterprise customers detect and resolve data-quality issues. My role combines growing the engineering organization and guiding technical decisions with hands-on development of agent workflows, context enrichment, model orchestration and sandboxed code execution.
Data engineering & production systems #
At 2Xideas, I built and maintained data infrastructure and production software in a regulated investment environment. My work spanned market-data integration, backtesting and internal packages for quantitative research, including turning research code into tested, maintainable production systems.
Applied machine learning #
At anacision, I developed automotive predictive-maintenance models using terabytes of data and optimization software for pharmaceutical personnel planning. I also implemented and tuned different GANs to generate differentially private synthetic data.
Predictive analytics for banking #
At msg for banking, I helped banks identify potential customers for loans, credit cards and investment products through predictive analytics and segmentation. My work also covered customer retention, churn analysis, financial time-series forecasting, backtesting and model validation.
Selected writing #
AI & engineering practice
What Engineers Contribute When AI Writes the Code
Problem framing, evidence and judgment when agents handle implementation.
From Code Review to Evidence Review
Evaluating runtime evidence and engineering decisions alongside the final diff.
Building the Systems That Produce Software
Context, system access and verification for reliable work with coding agents.
Architecture & implementation
From standard to fallback: the completions API in 2026 and what's next
Designing model routing and gateways as provider APIs diverge.
Getting started with Bifrost: why it's worth knowing about
Running an LLM gateway locally, with routing and governance across providers.
FastAPI's Dependency Injection Makes DB Access Effortless — and That's the Problem
How dependency injection can exhaust database connection pools, and how to prevent it.
Contact #
For a conversation about AI architecture, technical direction or implementation: