Senior Machine Learning Engineer (AI & Computer Vision)

Nearmap

1d•Remote

About The Position

This role is focused on translating research and development from the Nearmap AI & Computer Vision team into actionable data and insights. The Senior Machine Learning Engineer will be a key contributor to the algorithmic systems responsible for creating conflated Map Data products. Conflation involves integrating multiple data observations about the world into a unified map, prioritizing usability and a clear representation of real-world objects. Input data includes aerial imagery from various sources and times, as well as other geospatial data like property and permit information. A significant challenge is ensuring data provenance and accuracy while simplifying complexity for users. The team designs, builds, and operates software systems that process petabytes of data to extract insights from aerial imagery, utilizing Python scientific libraries, deep learning technologies like PyTorch, and GIS tools such as GEOS, Shapely, and GeoPandas/GeoPolars. Development is primarily in Python, with occasional use of compiled libraries for performance-critical tasks. Agentic coding tools are integrated into the workflow to accelerate exploratory and iterative development.

Requirements

Demonstrated 5+ year history working in a numerical field (e.g., applied mathematics, physical sciences, computer vision, geospatial analysis).
Ability to produce high-quality production numerical code with a focus on business outcomes.
Proficiency in scientific Python.
Experience using a Linux environment.
Experience with Git for source control.
Understanding of core ML concepts including regularization, hyperparameter optimization, and validation methods.
Ability to follow the scientific method of formulating hypotheses and applying statistical tests.
Adherence to best practices in modern software engineering for building robust, scalable machine learning systems.
Strong systems thinking approach, combined with pragmatism.
Commitment to software engineering principles for scientific Python, with a focus on clean code, robustness, and correctness.

Nice To Haves

Experience working with large datasets that do not fit into memory and require multi-node computation.
Experience working in a cloud-native environment using highly scalable compute.
Experience operationalizing numerical applications and workflows.
Experience with machine learning problems applied to image and geospatial data.
Tertiary qualifications in a field related to numerical science (Bachelor's degree in computer science, engineering, statistics, physics, etc.).

Responsibilities

Translate R&D from the Nearmap AI & Computer Vision team into data and answers.
Contribute to algorithmic systems for creating conflated Map Data products.
Combine multiple data observations about the world into a single, cohesive map.
Design, build, and operate software systems that transform aerial imagery to insight.
Profile conflation pipelines that produce unexpected artifacts.
Pair with CV researchers to understand the implications of model confidence scores for data quality.
Review code for new spatial indexing approaches.