Experience
Senior Machine Learning Engineer — NVIDIA
February 2024 – Present
Part of the production team of Alpamayo, a chain-of-thought Vision-Language-Action (VLA) model for end-to-end motion planning built on a state-of-the-art LLM backbone (LLaMA/QWEN), now deployed on Mercedes-Benz vehicles.
Key contributions: led the VRU safety ODD end-to-end (improving collision handling from 50% to 91%); main contributor to the 0.5B transformer backbone transition and primary developer of the trajectory decoder; established the first stable pre-training configuration across 6+ production model releases; developed novel control signal conditioning and data augmentation strategies delivering collision reductions of 15–50%; improved chain-of-thought reasoning capabilities; and developed an open-loop RL approach for trajectory refinement in rare and safety-critical scenarios.
Senior Engineer, Tech Lead / Team Lead — Motional Ad. Inc
February 2022 – January 2024
Led a team of nine researchers and engineers in the Motion Planning Research group, developing ML-based motion planning and decision-making algorithms for autonomous driving. Grew the team through two promotions, and one of our products led to the formation of a new team. Submitted two patents — one on learning-based intention prediction and one on motion planning for unstructured environments.
Developed the first ML trajectory generator model at Motional, improving planning performance by 10%, and contributed to Motional’s continuous learning framework, achieving a 15% average prediction performance improvement across several metrics.
PhD Candidate — Delft University of Technology (TU Delft)
Department of Cognitive Robotics | January 2018 – October 2022
Doctoral research on data-driven motion planning for autonomous navigation among humans, combining optimization-based techniques (Model Predictive Control) and learning-based techniques (Deep Reinforcement Learning, trajectory prediction). Thesis: Interaction-Aware Motion Planning in Crowded Dynamic Environments. Supervised by Prof. Javier Alonso-Mora.
Research Assistant — Fraunhofer IPA, Department of Robot and Assistive Systems
March 2016 – December 2017
Research on control theory and motion planning for mobile manipulators, applied to manipulation tasks and human-robot interaction.
Published in the International Conference on Computer Vision Systems 2019.
Graduate Trainee — European Space Agency (ESA)
Department of Guidance, Navigation and Control | April 2014 – February 2016
Research and development of guidance, navigation and control algorithms for autonomous spacecraft rendezvous and landing. Hardware-in-the-Loop simulations using robotic manipulators.
