Machine Learning Intern, Solution Enablement

Machine Learning Intern, Solution Enablement

Welcome to Planet. We believe in using space to help life on Earth.

Planet designs, builds, and operates the largest constellation of imaging satellites in history. This constellation delivers an unprecedented dataset of empirical information via a revolutionary cloud-based platform to authoritative figures in commercial, environmental, and humanitarian sectors. We are both a space company and data company all rolled into one.

Customers and users across the globe use Planet's data to develop new technologies, drive revenue, power research, and solve our world's toughest obstacles.

As we control every component of hardware design, manufacturing, data processing, and software engineering, our office is a truly inspiring mix of experts from a variety of domains.

We have a people-centric approach toward culture and community and we strive to iterate in a way that puts our team members first and prepares our company for growth. Join Planet and be a part of our mission to change the way people see the world.

Planet is a global company with employees working remotely world wide and joining us from offices in San Francisco, Washington DC, Germany, Austria, Slovenia, and The Netherlands.

About the Role: Deep Learning & Embedding-Based Approaches for Change Detection

Planet's Solution Enablement (SE) is seeking a Machine Learning (ML) intern to support the development of innovative land surface change monitoring systems based on AI solutions and satellite imagery. The successful candidate will develop scalable models that transform global Earth observations into actionable insights for the Defense and Intelligence, Civil Government, and Commercial sectors.

The challenge

The performance of standard pixel-based change detection methods is often reduced by the impact of vegetation seasonal cycle or atmospheric effects, leading to frequent misinterpretation. You will bypass these limitations by developing high-accuracy monitoring solutions for mission-critical applications, such as:

• Land parcel identification systems to help agricultural ministries automatically detect discrepancies between declared land use and actual ground truth (e.g., urbanization of arable land, deforestation, or illegal construction).
• Site Monitoring to support Defense and Intelligence by identifying structural temporal changes at specific sites, such as occupancy detection and classification.

Internship objectives

The primary goal is to build and evaluate a robust change detection model that distinguishes structural changes from natural seasonal variations and environmental effects. Key focus areas include the detection and classification of meaningful structural changes (e.g., construction of a road, occupancy changes, removal of a vineyard) over areas of interest.

The internship work will investigate whether embedding-based approaches (representing satellite imagery as dense vector embeddings) provide superior accuracy over traditional classification methods when detecting discrepancies between "reference" and "current" states.

You will join a passionate and innovative team that values collaboration and fosters a culture of continuous learning and impact-driven engineering. The team in which you will be working is distributed remotely across Europe with offices in Ljubljana and Haarlem.

This is a full-time, hybrid role which will require you to work in our Haarlem office at least 3 days per week.

Impact you'll own

• Literature Review: Investigate state-of-the-art methods for semantic change detection, specifically focusing on foundation models and image embeddings (e.g., using temporal encoders to suppress seasonal noise).
• Data preparation: Curate and process multi-source datasets, including Planet's imagery (Basemaps, Analysis Ready PlanetScope) and high-resolution aerial imagery, to build a robust training and testing pipeline.
• Algorithmic Innovation: Develop and test embedding-based workflows to classify changes into semantic categories (e.g., vegetation to built-up, arable to abandoned), or to detect new objects.
• Performance Benchmarking: Validate the prototype against existing baseline methods, specifically measuring its success in reducing false positives triggered by spectral similarities or natural cycles.

What you bring

• Bachelor's or Master's degree in Applied Mathematics, Computer Science, Remote Sensing or a related field
• Proficiency with Python and machine learning frameworks like TensorFlow or PyTorch
• Knowledge of remote sensing fundamentals and experience working with different types of remote sensing data
• Ability to be hands-on, be self-motivated, and think creatively
• Excellent technical communication, documentation and presentation skills
• Flexibility, proactivity, and curiosity in undertaking new tasks

What makes you stand out

• Experience in object-based detection methods
• Experience with deep learning at scale in a geospatial and/or remote sensing context

Application deadline: March 2nd, 2026 by 23:59 CET (Central European Time)

Benefits while working at Planet

• Paid time off including vacation, holidays and company-wide days off
• Employee Wellness Program
• Home Office Reimbursement
• Monthly Phone and Internet Reimbursement
• Tuition Reimbursement and access to LinkedIn Learning
• Equity
• Volunteering Paid Time Off

Why we care so much about belonging

We're dedicated to helping the whole Planet, and to do that we must strive to represent all of it within each of our offices and on all of our teams. That's why Planet is guided by an ultimate north star of Belonging—dreaming big as we approach our ongoing work. If this job intrigues you, but you're thinking you might not have all the qualifications, please... do apply! At Planet, we are looking for well-rounded people from around the world who can contribute to more ways than just what is listed in this job description. We don't just fill positions, we aspire to fulfill people's careers, most excited about folks who are motivated by our underlying humanitarian efforts. We are a few orbits around the sun before we get to where we want to be, so we hope you're excited to come along for the ride.