ML Engineer

This is a fully remote position, open to applicants in United States.

📋 Description

• Develop, customize, and implement foundation model-based deep learning systems that assess forest structure metrics from remotely sensed data.

• Fine-tune and adapt geospatial foundation models to serve as backbones for bespoke deep neural network heads.

• Incorporate trained models into Vibrant Planet’s automated production workflow.

• Sustain the surrounding data infrastructure.

• Prepare, curate, and manage training datasets sourced from satellite, lidar, and field plot data.

• Assess model performance using remote sensing accuracy metrics and ground validation data.

• Assist in experiment design, hyperparameter tuning, and ablation studies.

• Integrate machine learning models into automated geospatial workflows as containerized and orchestrated inference services.

• Develop and maintain STAC infrastructure for model input and output discovery, cataloging, and access management.

• Design and implement Airflow DAG pipelines that ensure idempotency, observability, and fault tolerance.

• Oversee data ingestion, preprocessing, and quality control processes.

• Monitor pipeline performance and model drift; establish alerting and automated retraining protocols.

• Create model cards summarizing modeling techniques and outcomes.

• Write and contribute to scientific publications.

• Bridge SciDev, Data Engineering, and Product by translating requirements and aligning objectives.

• Document pipelines, architectures, and operational processes.

• Engage in code reviews, architectural discussions, and sprint planning sessions.

• Adhere to information security, secure development practices, change management, and customer data protection protocols.

⛳️ Requirements

• M.S. in Computer Science, Machine Learning, Remote Sensing, Data Science, Ecology, or a related quantitative discipline, or equivalent professional experience.

• Over 3 years of experience in developing, training, and deploying deep learning models, with PyTorch being preferred.

• Strong proficiency in Python, including libraries such as NumPy, pandas, xarray, and scikit-learn.

• More than 3 years of experience with geospatial data processing tools like rasterio, GDAL, geopandas, and shapely.

• Experience in building and maintaining data pipelines using Airflow, Prefect, Dagster, or similar frameworks.

• Proficient with Git, GitHub, code review processes, and CI/CD practices.

• Familiar with Docker and cloud platforms, with a preference for AWS.

• Knowledge of STAC specifications and geospatial data catalog infrastructure.

• Excellent written communication skills with the ability to contribute to scientific publications and technical documentation.

• Basic understanding of forest ecology, remote sensing concepts, or natural resource science.

• Capability to work collaboratively within interdisciplinary teams.

• Self-motivated, able to manage time effectively, and work independently in a remote-first setting.

• Ability to communicate effectively between scientific and engineering audiences.

• Must be authorized to work in the U.S. without requiring visa sponsorship.

• Preferred qualifications: Ph.D.; experience with geospatial foundation models and self-supervised learning; knowledge of Kubernetes and distributed computing; familiarity with MLflow or W&B; experience with PostgreSQL/PostGIS and message queues; relevant publications.

🏝️ Benefits

• Health, dental, and vision insurance.

• 401(k) plan.

• Unlimited PTO policy.

• Company equity.

• Cell phone stipend (per pay period).

• One-time home office setup allowance.