Computer Vision & Machine Learning Engineer
Buzz Solutions3 hours ago
Remote, United StatesMid Level / Senior
H1B Sponsor
Responsibilities
- Own computer vision projects from problem framing through production deployment and monitoring.
- Develop solutions for equipment defect detection, thermal anomaly identification, vegetation encroachment monitoring, and intrusion surveillance.
- Translate client requirements and raw data into working computer vision solutions.
- Evaluate and implement methods from ML and computer vision research papers.
- Design experiments with hyperparameter tuning, ablation studies, baselines, and structured error analysis.
- Select and adapt model architectures based on latency and accuracy requirements.
- Build Python libraries, data pipelines, annotation workflows, model-serving pipelines, and quality-monitoring systems.
- Manage experiment tracking, dataset and model versioning, configurations, environments, and checkpoints.
- Conduct code reviews, write integration tests, and uphold software quality practices.
- Communicate research findings, technical decisions, and model limitations to teammates, stakeholders, and clients.
Requirements
- 2–5 years of industry experience in computer vision and machine learning.
- Strong knowledge of object detection, semantic segmentation, image classification, vision transformers, foundation models, vision-language models, and similarity search.
- Experience taking at least one machine learning model into production and maintaining it.
- Experience selecting, fine-tuning, and adapting CNN, transformer, and foundation-model architectures.
- Ability to read ML research papers, implement key ideas, debug training instabilities, and perform systematic error analysis.
- Proficiency in Python, PyTorch, Lightning, OpenCV, NumPy, pandas, Scikit-Learn, FastAPI, and Pydantic.
- Strong software engineering practices involving Git, Pytest, GitHub Actions, Docker, reproducible environments, experiment tracking, model versioning, ML DevOps, and Python type hinting.
- Ability to independently own technical projects from problem framing through production deployment.
- Preferred experience with multimodal computer vision, custom object detection, generative models for data augmentation, GIS or drone-metadata-enriched imagery, model quantization, edge latency optimization, and large-scale hyperparameter tuning.
- Experience in energy, utilities, geospatial, or industrial inspection domains is desirable.
- United States work-authorization sponsorship is not provided.
Categories
About Buzz Solutions
Buzz Solutions provides a platform for teams to manage and analyze data, collaborate, and export inspection results, fostering smart, stable, and resilient infrastructure inspections. We automate the process of infrastructure inspections for faults and anomalies by analyzing millions of visual data points captured by helicopters, drones and linemen in the field. Using our solution, our customers are saving immense time and money as a part of their inspections, while drastically improving the efficiency of their infrastructure inspections.