Raphael Labaca-Castro

Timeline

1. 2004–2016

   Studies in Argentina and Bochum

   He studies information systems at Universidad Tecnológica Nacional in Argentina from 2004 to 2011, with an exchange stay at the Institute for Neural Computation at Ruhr-Universität Bochum in 2009–2010. From 2012 to 2016, he earns a master's in information security at Universidad de Buenos Aires; his thesis examines how adversaries could compromise genomic FASTA files.

2. 2011–2017

   Research lead at ESET

   From April 2011 to September 2017, he is research lead at ESET, first leading a regional group of researchers in Argentina and then relocating to Germany to manage a global team. He also writes for WeLiveSecurity, including the August 2017 article “Malware coded into synthetic genomes”.

3. 2017–2022

   PhD at Research Institute CODE

   From 2017, he works toward a PhD in adversarial machine learning at the Universität der Bundeswehr München, graduating summa cum laude with distinction. The research takes place at Research Institute CODE in Munich under Gabi Dreo Rodosek of the Universität der Bundeswehr and Lorenzo Cavallaro of University College London, and examines how machine learning models respond to carefully crafted adversarial manipulations.

4. 2019–2021

   Adversarial attacks on malware classifiers

   In 2019, he publishes “Attacking Malware Classifiers by Crafting Gradient-Attacks that Preserve Functionality” at ACM CCS and AIMED at IEEE TrustCom, which uses genetic algorithms to evade malware classifiers. In February 2021, he and five coauthors post “Realizable Universal Adversarial Perturbations for Malware” to arXiv, a paper that explores universal adversarial perturbations in the context of malware classification. AIMED-RL, written in 2021 with Sebastian Franz and Gabi Dreo Rodosek for ECML PKDD, uses a Distributional Double Deep Q-Network agent to generate adversarial examples that make models misclassify malware without compromising its functionality.

5. 2020–2022

   Residency at Google X

   He is a PhD resident at Google X in Mountain View, California, from February to May 2020, and reaches up to a tenfold performance improvement on a real-world problem. From February 2021 to February 2022, he is an ML engineer at Google X in Munich, where he launches a project to support crypto-agility within organizations.

6. February 2022

   Machine learning lead at SandboxAQ

   In February 2022, he becomes machine learning lead at SandboxAQ in Munich, leading a team where AI and quantum disciplines meet. He is among the early founding members of SandboxAQ, a spin-off company that integrates artificial intelligence with quantum technologies.

7. 2023

   A book on adversarial malware

   Springer Nature publishes Machine Learning Under Malware Attack in 2023, a book based on his dissertation. In December 2023, he coauthors “Secrecy and Sensitivity: Privacy-Performance Trade-Offs in Encrypted Traffic Classification” with Spencer Giddens, Dan Zhao, Sandra Guasch, Parth Mishra and Nicolas Gama for the MLSystems workshop at NeurIPS 2023. The paper applies differential privacy to machine learning models that classify encrypted and non-encrypted network packets.

8. As of October 2026

   Staff engineer on AQMed

   He is a Staff Machine Learning Engineer on the AQMed team at SandboxAQ, where he works on detecting cardiovascular diseases with a magnetocardiography device. The FAME library, which he authors, assesses ML-based classifiers against adversarial attacks with malicious software. He also guides early-stage startups in defining AI products.