Dr.-Ing. Ee Heng Chen

Summary

• Built Large Language Model (LLM) and Generative AI (GenAI) systems that turn unstructured regulatory and technical documents into structured, reviewable outputs, with Retrieval-Augmented Generation (RAG) and evaluation pipelines.
• Rolled out n8n to production as the company-wide automation platform, now used by ~10 engineers across 40+ workflows.
• Configured MLOps toolchains (DVC, MLflow, AWS) for reproducible and auditable ML workflows.
• Personal projects across the AI agent stack (multi-agent orchestration, Model Context Protocol (MCP) tooling, replay testing) on one principle: the LLM classifies, a deterministic policy layer decides and writes the audit trail.
• Dr.-Ing. in computer science, with delivery experience across automotive, clinical and industrial settings.

Experience

1. VAIVA GmbH

   12/2023 – Present

   Machine Learning Engineer

   • Conceived and built an agentic GenAI application, working cross-functionally with the requirements engineering team, that ingests regulatory documents and drafts structured system requirements, so engineers review a draft instead of starting from a blank page. Designed the architecture, expert-authored agent skills, deterministic format checks and the Claude Code agent harness integration.
   • Deployed n8n to production as the company automation platform and led adoption across ~10 engineers, with 40+ workflows in use and a shared workflow library versioned in GitLab.
   • Designed and delivered an LLM-based internal testing tool using LangChain and Gradio, incorporating RAG, structured output parsing, prompt engineering and evaluation pipelines; led it from proof of concept to functional prototype.
   • Integrated AWS Bedrock as the LLM backend, giving applications API access to cloud-hosted foundation models.
   • Set up MLOps toolchains for model training (data versioning with DVC, experiment tracking with MLflow) and integrated them into the company cloud environment with the DevOps team.

2. Munich Institute of Robotics and Machine Intelligence

   04/2021 – 08/2023

   Postdoctoral Researcher · Machine Learning

   • Coordinated researchers and clinical stakeholders at the German Heart Center Munich, and authored two successful ethics proposals for patient data collection.
   • Modeled patient and clinician actions in intensive care unit (ICU) settings using graph convolutional networks (GCNs) in PyTorch.
   • Built the supporting data pipeline: an ICU data-capturing system on Raspberry Pi and Intel RealSense devices, plus a Django web application for visualization.

3. BMW Group

   09/2017 – 03/2021

   Doctoral Researcher · Machine Learning / Computer Vision (ADAS)

   • Developed a vision-based system determining whether a traffic junction is safe to cross for driver assistance and self-driving vehicles; trained and evaluated CNNs in TensorFlow.
   • Used Docker and Kubernetes to manage training and evaluation jobs on compute clusters, and maintained the compute infrastructure for research and prototyping.
   • Published 5 conference papers; supervised 2 interns and 4 student theses.

Projects

Education

• Technical University of Munich

  09/2017 – 07/2022

  Dr.-Ing. (Ph.D.) Computer Science, Department of Informatics

  Thesis: Development of Vision-based Driving Strategies for Traffic Junction Scenarios.

• Technical University of Munich

  04/2015 – 08/2017

  M.Sc. Robotics, Cognition, and Intelligence, Department of Informatics

• Heilbronn University of Applied Sciences

  04/2011 – 03/2015

  B.Eng. Mechatronics and Microsystems, Faculty of Mechanics and Electronics

Skills