Lance DiAngelis

Experience

• 

  DynCorp International

  Jun 2008 - Aug 2011

  Aircraft Mechanic Helper

  - Coordinated emergency response to catastrophic in-flight valve failure, preventing fuel spillage into Clean Water Act protected waterway.- Troubleshot and repaired C-130 (J,H,R,T models) aircraft engines with scheduled and unscheduled repairs to ensure flight readiness in support of NAS Patuxent River test and evaluation missions.- Conducted daily inspections of critical aircraft components pre/post flight.

• 

  Affinity Consultants, Inc.

  Feb 2013 - Jul 2015

  Environmental Scientist

  • Evaluated environmental programs and implemented corrective actions for a large-scale railroad operation with oil storage exceeding 1 million gallons.• Conduct annual evaluation of client Standard Operating Procedures and recommend corrective actions to enhance site environmental, health, and safety programs.• Collected field sensor data, maintained sampling database, and developed long term time-series of sampling data for 20 clients.

• 

  University of North Dakota

  Aug 2015 - Apr 2021

  • Acquired, preprocessed, and cleaned large image and tabular datasets, including 35 years of Landsat satellite imagery and USGS environmental data, using Python.• Developed robust image feature extraction, time-series, and multi-variate machine learning models to leverage multi-modal datasets.• Owned reporting and analysis of the UND Greenhouse Gas report findings; assisting in the development of corrective actions for meeting established KPI’s.• Designed and managed GIS geodatabase to create informative maps of oil spill clusters in Western North Dakota. Show less

  • Pre-Award Officer

    Sept 2019 - Apr 2021

  • Graduate Research Assistant

    Aug 2015 - Aug 2019
• 

  Freelance

  Apr 2021 - now

  Freelance Computer Vision Engineer

  • Investigate the training of convolutional neural networks (VGG16, GoogLeNet, ResNet-50, etc.) from the ground up to replicate SOTA results on ImageNet, with a focus on enhancing model efficiency through improved GPU and memory utilization.• Fine-tune and deploy end-to-end object detection (YOLOv3/4, SSD) and semantic segmentation (UNet, SegFormer) networks using NVIDIA DALI, TAO Toolkit, TensorRT, and Triton Inference Server.• Develop custom Jetson Nano and Raspberry Pi computer vision projects with the Raspberry Pi v2 camera.• Deploy TensorFlow Lite quantized models for computer vision tasks on edge devices, resulting in 4x inference throughput.• Manage Docker and NVIDIA NGC Containers for deploying computer vision projects.• Collect, label, and annotate custom image dataset for fine-tuning object detection models.• Implement image processing and data augmentation pipelines to improve the accuracy and performance of computer vision models. Show less