Engineer with 4 years of experience across academia and industry, developing AI and CV based systems with a focus on optimized inference, versioned training cycles, and reproducible data pipelines.
- AI Engineer @ Indra (2025 - present): defect detection on scanned documents (YOLO) and structured data extraction from PDFs (DocLayout-YOLO + visual embeddings)
- CV Engineer @ LABINM - UPAO (2024 - present): real-time blueberry detection and counting (YOLO + BoT-SORT, mAP50 0.87), optimized to 16 ms inference with TensorRT FP16 on Jetson Xavier
- AI Implementation Engineer @ YaVendio (2024 - 2025): LLM sales agents over real customer conversations in production (LangGraph)
- Jr. Full Stack Developer @ IDEGO (2024): backend microservices for Maersk Peru's order management platform (Django on Azure)
- LabelFlow: open-source video annotation tool with optical-flow label propagation
- mlops-detection: MLOps framework to train and evaluate YOLO detectors (v8-v11) with YAML-defined experiments and hyperparameter sweeps
- mlops-blueberry-counting: YOLO detector + tracking pipeline for counting blueberries in video
- mlops-classification-blueberry: ripeness classification with a versioned training pipeline
- vslam-notebooks: visual SLAM experiments and notes
Python C++ PyTorch YOLO Detectron2 ONNX TensorRT ROS AWS FastAPI