kaiyin.zhong

Career

Experience

From bedside medicine to terabyte-scale genomics to production ML platforms to AI infrastructure. Each layer still carries the ones beneath it.

  1. 2025 — now

    Swoep — Chief Data Officer

    Designing and running Swoep's AI infrastructure from the ground up: a secure, scalable data foundation for LLM-powered agents, with co-ownership of architecture decisions and technology selection, shoulder-to-shoulder with the CTO.

    • LLM agents and MCP servers for SQL/Python execution and third-party API integration
    • An LLM load balancer that brought token-quota downtime to zero, plus CI/CD pipelines for LLM applications
    • Platform migrations — MSSQL to PostgreSQL, Azure to Hetzner, Azure OpenAI to AWS Bedrock — cutting infra costs by thousands of euros a month
    • Set up and ran the ISMS that took Swoep to ISO 27001 certification

    LLM agentsMCPPostgreSQLAWS BedrockISO 27001

  2. 2019 — 2025

    Gemvision — Senior Data/ML Engineer & System Architect

    Progressed into a leadership role with end-to-end ownership of the data platform and infrastructure, serving healthcare and enterprise partners. Advised executive leadership on emerging technologies and helped colleagues adopt AI-assisted development.

    • Architected an MLOps platform for parallel model development: real-time on-device inference plus terabyte-scale batch processing
    • Productionized ML models for clinical decision support; APIs served predictions and patient data to EHR systems for 10k+ monthly active healthcare professionals
    • ETL pipelines standardizing 5M+ annual medical records into a unified feature store for analytics, reporting, and model training
    • Zero-downtime migration of a legacy system to Azure with Terraform and Kubernetes (AKS) for critical healthcare applications
    • GenAI developer tooling: an LLM-powered CLI in Rust automating documentation and hour registration, and a chat interface letting execs query complex datasets without code
    • CI/CD with GitHub Actions and ArgoCD cut release cycles from half an hour to minutes; Grafana + Prometheus monitoring shortened incident resolution
    • Key role in ISO 27001 and NEN 7510 certification; mentored interns and junior data engineers

    MLOpsAKSTerraformKubernetesPostgreSQL

  3. 2017 — 2019

    Developers.nl — Machine Learning & Data Engineer

    Consulting engineer on industrial ML and data projects.

    • Data pipelines processing terabytes of sensor and imagery data for maritime and oil clients
    • Image segmentation models trained and deployed in industrial environments, on both edge and cloud
    • Scala APIs serving ML model predictions with sub-second response times
    • Infrastructure with Docker and Proxmox, DevOps pipelines in Azure DevOps, secure payment integration with Mollie

    Scalacomputer visionDockerAzure DevOps

  4. 2012 — 2017 · Education

    Erasmus University — PhD, Computational Genomics

    Built large-scale pipelines analysing multi-terabyte genomic datasets — where engineering took over from medicine for good.

    genomicspipelinesRPythonstatisticsregression

  5. 2009 — 2012 · Education

    Peking University — MSc, Neuroscience

    Focus on data analysis and computational modelling.

    data analysiscomputational modelling

  6. 2006 — 2009

    Aviation Industry Hospital — Resident Physician

    Clinical practice — first hands-on experience with structured data in healthcare.

  7. 2001 — 2006 · Education

    Wuhan University — Bachelor of Medicine