At Varada, we don’t just apply AI. We build with it. From predictive analytics to machine learning and deep learning, AI is embedded across everything we deliver. JIVA360 and BEDDE are powered by these capabilities, supporting real-time insight, automation, and mission-readiness. That same innovation is driving our impact in Geospatial Intelligence, where our AI work is transforming how agencies like NASA analyze data and make decisions.
Melanie Frost, our Remote Sensing Imagery Analyst on the NASA Advanced Computing Services (NACS) contract, brings this mission to life. Based at Goddard Space Flight Center, she supports the NASA Center for Climate Simulation and leads groundbreaking work in Earth and planetary sciences using AI-powered GEOINT.
Mapping Uncharted Territories
The Goddard Space Flight Center in Greenbelt, Maryland, is a hub for groundbreaking research, housing the highest concentration of Earth scientists in the world. Melanie and her team are engaged in mapping previously uncharted territories using advanced satellite data. One of their notable projects involved the ICESat-2 satellite, launched in 2018 to monitor changes in Earth's ice. This satellite, which fires 10,000 lasers per second, enables scientists to estimate tree canopy heights across North America's boreal forests. Utilizing a machine learning algorithm known as Random Forest, Melanie’s team built a model from approximately 30 billion tree canopy height data points and climatological data to predict how the North American Boreal Forest will shift over the next century.
“ICESat-2's data allows us to estimate how tall the trees are in an area by measuring the difference between laser points that bounce off the earth and those that bounce off the tree canopies. This helps us understand changes in forest structure over time,” Melanie explains.
Harnessing AI and Deep Learning for Climate Science
Artificial intelligence (AI) and machine learning (ML) are integral to Melanie's work. Her team has made initial strides in developing a pre-trained model for the MODIS satellite surface reflectance process, which has been collecting data since 2000. By training vast amounts of data on supercomputers, they aim to create models that scientists can query for immediate insights, without requiring the data labelling and compute resources. This initiative aligns with NASA's commitment to open science, making data accessible to policymakers and researchers worldwide.
"The Data Science Group is building a foundation model for the MODIS satellite, which orbits the Earth every day, collecting a huge amount of data. By training this data on our supercomputers, we can provide immediate insights for scientists, much like how ChatGPT operates," Melanie explains.
Tackling Global Challenges with AI
Melanie's team also employs AI for land cover classification projects, such as identifying crop types in Senegal to assist with agricultural planning. AI has the potential to revolutionize fields like climate simulation, enabling scientists to classify and interpret large spatial areas of satellite data with unprecedented accuracy. This technology extends to mapping celestial bodies, where applying AI and Machine Learning techniques helped to identify five times more craters on Saturn's moons than previously known.
"One of our projects in Senegal involves using AI models to classify land types. This can help decision-makers plan for future land use and agricultural needs, especially in the face of climate challenges," says Melanie.
The Power of Supercomputing
At the heart of NCCS is the Discover supercomputer, boasting 8,000 trillion floating-point operations per second. This immense computational power enables simulations ranging from short-term weather forecasts to long-term climate projections. One remarkable achievement involved simulating a black hole, which would have taken a decade on a laptop but was completed in just two weeks using Discover.
"The Discover supercomputer allows us to run complex simulations that would otherwise be impossible. For example, a researcher recently simulated a black hole, which would have taken ten years on a traditional laptop, but with Discover, we did it in just two weeks," Melanie notes.
Ensuring Ethical AI
While AI presents exciting opportunities, Melanie emphasizes the importance of ethical considerations. Proprietary data and potential biases in AI models are challenges that NASA addresses through in-house capabilities and stringent ethical standards. By ensuring unbiased data and transparent models, NASA aims to leverage AI responsibly.
"AI is only as good as the data it's trained on. We need to be vigilant about biases in our data and models to ensure ethical and accurate outcomes," Melanie highlights.
A Dream Job. A Bigger Mission.
Melanie, a data scientist by trade, transitioned from the Air Force, where she specialized in exercise science and human performance analytics. With Varada Consulting for a year and a half, she appreciates the collaborative environment and the opportunity to work with world-class scientists.
"This is really my dream job. Working with people who are not afraid of big challenges and who want to save the world is incredibly rewarding. NASA is exactly as fun as you think it is," Melanie shares enthusiastically.
Inspiring the Next Generation
Melanie actively mentors and encourages future scientists and engineers. Her advice for aspiring NASA talent:
• Explore programs like OSTEM, Pathways, and DEVELOP (which she completed)
• These opportunities exist from high school to postdoc levels
• Follow NASA’s pages and social media to track openings
For those interested in data science or AI, Melanie emphasizes three key skills: computer programming, math, and curiosity. "This field is all about continuous learning. We are always taking classes and improving. Finding a mentor can help you navigate the plethora of resources available online. Never stop learning and taking classes; that is crucial," she advises.
The Future of AI at NASA
Melanie is optimistic about the future of AI, envisioning it as a tool to eliminate mundane tasks and enhance creative problem-solving. AI's potential to uncover unknown solutions and communicate previously unconsidered data is promising. However, ethical implementations and continuous learning remain paramount to harnessing AI's full potential.
“AI has the potential to transform our work by eliminating the drudgery of administrative tasks, allowing us to focus on creative problem-solving. It’s an exciting time to be at the forefront of this technology,” she says.
Innovating with AI and GEOINT at Varada
Melanie’s contributions to NASA are a reflection of the culture at Varada. We build solutions that advance discovery, support real missions, and reshape how federal agencies use data.
But her achievements are just the beginning. At Varada, our AI and GEOINT capabilities power strategic planning and operational readiness across government. From forecasting trends in satellite data to optimizing DevSecOps with deep learning models, we continue to lead by building what others haven’t yet imagined.
Looking to integrate real AI into your mission?
Reach out at info@varadaconsulting.com
