ALEX GOLDHOORN

Professional Summary

Senior Data Scientist with 12+ years of industry experience specializing in logistics optimization, discrete event simulation, and LLM agent systems. Built production systems serving 120,000+ couriers across 23 countries. PhD in Robotics with expertise in planning under uncertainty.

Core Skills

Professional Experience

Artificial Intelligence Engineer • AstraZeneca (contract via Hybrid Mind) • Mar 2026 - Present

• Agentic Pipelines: Design and build LLM agent pipelines for enterprise data workflows using LangGraph, with LangSmith tracing
• Agent Evaluation: Define evaluation suites measuring agent accuracy, faithfulness, tool use, and cost across offline and online tests
• Semantic Modelling: Model a semantic layer and domain ontology/knowledge graph over enterprise data so agents query modelled entities and relationships, not raw tables

Senior Data Scientist • Meight (Freight Logistics) • Dec 2024 - Feb 2026

• LLM Agent Evaluation: Designed the evaluation framework combining classical ML metrics with LLM-as-a-judge evaluation (faithfulness, cost) for AI agents automating freight manager tasks, ensuring production reliability
• Tour Optimization: Developed algorithms integrating real-time telematics and EU driving-time regulations to optimize freight planning and reduce empty running, boosting Revenue Per Kilometer (RPK)

Senior Data Scientist • Glovo (On-Demand Delivery) • Nov 2019 - Nov 2024

• Delivery Simulator: Led architecture and refactor of discrete event simulator for Glovo's delivery network (23 countries, 120,000+ couriers), enabling rapid experimentation and hyperparameter optimization across the full delivery lifecycle
• Matching Optimization: Built automated optimization pipeline for matching cost function, balancing delivery speed, courier earnings, and operational efficiency at global scale
• Route Optimization (VRP): Implemented Vehicle Routing Problem solution for Live Multibundling, enabling dynamic order batching and real-time route optimization, reducing delivery times and costs

Senior R&D Software Developer • Atomian (Cognitive AI) • Jun 2017 - Nov 2019

• Built ETL pipelines and prediction models for a cognitive NLP platform extracting structured data from unstructured patient reports and legal documents

PhD Researcher • IRI - Institut de Robòtica (CSIC-UPC) • Apr 2011 - Jun 2017

• Thesis: Human search and tracking in urban environments using humanoid robots — POMCP, particle filters, Reinforcement Learning
• Validation: Conducted 3km+ of autonomous experiments with humanoid robots in real-world urban settings, handling uncertainty and noisy sensors

Education

Certifications

• Anthropic Education (8 courses, Apr 2026): Claude 101, Claude Code, Building with the Claude API, agent skills, subagents, MCP intro & advanced
• LangChain Academy (7 courses, Mar 2026): LangChain intro, LangSmith, reliable agents, observability & evals, LangGraph (ambient agents, deep research, deep agents)
• DeepLearning.AI (Nov 2025–Aug 2026): AI for Medicine (3-course professional cert), ML in Production, Generative AI for Software Development, PyTorch for Deep Learning, Agentic AI, and 46 short courses on agents, evaluation, RAG, MCP, transformers, voice AI, code review, AI coding workflows, and more
• Deep Learning Specialization (5-course) — DeepLearning.AI / Coursera (Jul 2019)
• Discrete Optimization — Georgia Institute of Technology / Coursera (Dec 2020)

Languages

Dutch (native), English (fluent), Spanish (fluent), Catalan (proficient)