PhD Student · Inria Chroma · Université Grenoble Alpes
AI-driven motion planning and decision-making for safe autonomous driving.
I'm a PhD student at Université Grenoble Alpes, supervised by Anne Spalanzani and Christian Laugier. I work in Inria's Chroma lab, where my research focuses on AI-driven algorithms for safe motion planning and driving decision-making for autonomous vehicles.
I hold an MSc in Mobile, Autonomous and Robotic Systems from Grenoble INP (2022), a Postgraduate Diploma in Artificial Intelligence (2021) and a BEng in Mechatronics Engineering (2018) from Universidad Autónoma de Occidente.
Presented “PlanFlow: Local trajectory planning for autonomous driving using flow-guided occupancy grids” at the 2026 IEEE Intelligent Vehicles Symposium (IV) in Detroit, USA.
Presented “TLCFuse: Temporal Multi-Modality Fusion Towards Occlusion-Aware Semantic Segmentation” at IEEE IV 2024 in Jeju, South Korea.
Started my PhD at Université Grenoble Alpes and Inria, supervised by Anne Spalanzani and Christian Laugier.
Presented “TransFuseGrid: Transformer-based Lidar-RGB fusion for semantic grid prediction” at ICARCV 2022.
Joined Inria's Chroma lab as a Research Engineer working on multi-modal sensor fusion for autonomous driving.
Defended my Master's thesis “Transformer-based Lidar-RGB Fusion for Semantic Grid Prediction in Autonomous Vehicles”, carried out at Inria's Chroma lab.
Started an internship at Inria's Chroma lab on sensor fusion for robust situational awareness in autonomous vehicles.
Started an MSc in Mobile, Autonomous and Robotic Systems at Grenoble INP.
Finished a Postgraduate Diploma in Artificial Intelligence at Universidad Autónoma de Occidente.
“High-level camera-LiDAR fusion for 3D object detection with Machine Learning” presented as a poster at the LatinX in AI Workshop at CVPR 2021.
Started a Postgraduate Diploma in Artificial Intelligence at Universidad Autónoma de Occidente.
“Application of Transfer Learning for Object Recognition Using CNNs” selected for an extended version in Springer CCIS vol. 833.
“Application of Transfer Learning for Object Recognition Using CNNs” published at IEEE ColCACI 2018.
Flow-guided occupancy grids augment bird's-eye-view occupancy maps with per-cell flow vectors, combining occupancy, motion and uncertainty in one input. PlanFlow plans on them with a compact CNN backbone and an MLP decoder, trained with L1 and collision-aware losses, and achieves low collision rates on nuScenes against BEV-input baselines.
Temporal cues make bird's-eye-view semantic grid segmentation more robust to occlusions. The architecture processes multi-step inputs, each one fusing LiDAR and camera readings.
A transformer fuses LiDAR data with RGB images from multiple cameras to predict semantic grids of the ego vehicle's surroundings.
3D vehicle detection on nuScenes with classic machine learning such as DBSCAN and SVMs.
Real-time recognition of laboratory equipment at Universidad Autónoma de Occidente, using Inception-V3 as a pre-trained CNN.
DeepLearning.AI & Amazon Web Services · Coursera
How LLM-based generative AI works across the model lifecycle: transformer architecture, prompt engineering, fine-tuning and RLHF, and deployment in real applications.
CredentialInstitut Mines-Télécom · Coursera
Credential
Implementation of A Neural Algorithm of Artistic Style with VGG-19.
LeNet-5 and VGG16 in TensorFlow on MNIST, trained on GPU in Colab.
Supervised (KNN, decision trees, ANN) and unsupervised (DBSCAN, PCA) methods, plus feature engineering to improve model performance.
A GAN that generates car images from CIFAR-10, trained on GPU in Colab.
OpenCV image processing and a KNN digit classifier; the recognised digits are drawn back into a grid matching the image.
