Senior AI – Computer Vision Engineer

Oxy•Houston, TX

1d•Onsite

About The Position

Oxy is seeking an experienced and innovative Senior AI – Computer Vision Engineer to join the AI Center of Excellence (ACE) group based in Houston, TX. This individual contributor role focuses on designing, developing, and deploying production‑grade computer vision solutions across Oxy, supporting a wide range of industrial, operational, and subsurface use cases. Oxy produces, markets and transports oil and natural gas to maximize value and provide resources fundamental to life. The company leverages its global leadership in carbon management to advance lower-carbon technologies and products. Headquartered in Houston, Oxy primarily operates in the United States, Middle East and North Africa. Oxy strives to attract and retain talented employees by investing in their professional development and providing rewarding opportunities for personal growth. Our goal is to meet the highest employer standards by ensuring the health and safety of our employees, protecting the environment and positively impacting our communities where we do business.

Requirements

6+ years of hands‑on experience in computer vision or applied deep learning
Excellent Python skills (required), including writing clean, efficient, production‑ready code
Strong experience with PyTorch, TensorFlow, CNN‑based architectures, transformers, and Vision‑Language Models
Ultralytics YOLO experience required, including training and tuning on real‑world datasets
Practical familiarity with Hugging Face and Roboflow
Experience working with GPU‑accelerated workloads and CUDA‑enabled deep learning frameworks
Experience developing or running ML workloads on AWS, including Amazon SageMaker and GPU instances
Strong experience working in Linux environments
Excellent teamwork, communication, and presentation skills, with the ability to explain complex technical concepts to both technical and non‑technical audiences
Demonstrated contributions to computer vision research, including peer‑reviewed publications, conference papers, or equivalent applied research output

Nice To Haves

PhD in Computer Science or a related technical field preferred, or equivalent industry experience building production computer vision systems

Responsibilities

Design and select appropriate computer vision model architectures for classification, detection, segmentation, and object tracking
Work with classification architectures such as ResNet, VGG, EfficientNet, and MobileNet, and segmentation architectures such as U‑Net
Build, train, fine‑tune, and optimize models using Ultralytics YOLO for object detection and segmentation
Develop deep learning models using PyTorch and TensorFlow
Lead research and development (R&D) efforts to evaluate, prototype, and adopt state‑of‑the‑art (SOTA) computer vision models and techniques where they provide business or operational value
Stay current with advances in computer vision research, including new architectures, training methods, and foundation models, and translate relevant innovations into practical solutions
Leverage Hugging Face for pretrained backbones, model assets, and rapid experimentation
Apply Vision‑Language Models (VLMs) to multimodal computer vision workflows (e.g., OCR, image‑to‑text, prompt‑driven visual understanding)
Design, manage, and continuously improve image and video labeling workflows, using Roboflow or similar annotation tools
Deliver computer vision models for surface and downhole image analysis, including lithology, facies, and textural interpretation
Optimize computationally heavy training and inference workloads, including GPU utilization, memory efficiency, and throughput/latency tradeoffs
Work with GPU‑accelerated environments (CUDA‑enabled frameworks) and AWS‑based ML infrastructure, including Amazon SageMaker when appropriate
Collaborate closely with cross‑functional teams (AI platform, software engineering, domain experts) and mentor junior engineers
Communicate technical findings, experimental results, and recommendations clearly through presentations, demos, and written documentation