Stories

Though he didn’t know exactly what kind, Ignacio López-Francos always dreamed of becoming an engineer. Inspired by his early love of science fiction—and following in the footsteps of his grandfather, a respected Spanish automotive engineer—he enrolled in the Industrial Engineering School at the Polytechnic University of Madrid (UPM). During his last year, supported by a scholarship, he moved to Chicago to study at Illinois Institute of Technology (IIT), where he earned his second master’s degree.

His first job out of school involved designing and managing the end-to-end supply chain for a quick-service restaurant startup. He helped the company grow from five locations in Illinois to 25 worldwide, evolving it into a multi-brand franchise. It was there that he first applied principles of efficiency, optimization, and automation learned in school. This experience compelled him to upskill in programming, data management, and machine learning—through online courses he took at night and on weekends, ultimately allowing him to build a custom semi-automated Enterprise Research Planning (ERP) system. Continuing education soon became a constant in his career:

“As I was striving to further optimize every process along the company’s supply chain, I saw machine learning and artificial intelligence (AI/ML) as a natural extension of my industrial engineering education. As I started coding and integrating models that used historical sales data, weather, seasonal variations, and other variables, I could see—in almost real time—the impact of improving our demand forecasting at the store and product levels.”

Ignacio next pivoted to a boutique consulting firm, building big data and data science solutions across industries like finance, asset management, healthcare, cloud computing, and even pest-control. In 2015, he joined United Airlines and became a founding member of the Data Sciences & Innovation group, where he led AI R&D projects such as a travel advisor chatbot, a destination recommendation engine, and a smart monitoring system that leveraged airport camera feeds to optimize turnaround operations.

He eventually moved to the Bay Area to work for Facebook, supporting business and operations teams in developing and scaling Facebook’s marketing partner ecosystem. During that NeurIPS event, Ignacio learned about NASA’s Frontier Development Lab (FDL)—an AI research accelerator supported by Google and NVIDIA—and he joined the two-month program as a mentor the following summer. His work at FDL caught the attention of NASA sponsors, who then offered him a role to join the agency full-time to help assess their entire science portfolio for opportunities where AI could accelerate mission outcomes.

One of Ignacio’s proposals, which received seed funding from FDL through a NASA Cooperative Agreement, addressed a critical knowledge gap in NASA’s Artemis campaign to return humans to the Moon: obtaining high-resolution imagery of permanently shadowed regions (PSRs) near the Moon’s South Pole, where “trapped” water-ice is believed to lie in cold, dark craters. In collaboration with researchers from Oxford University, ETH Zurich, and the Luxembourg Space Agency, his team developed a physics-based AI algorithm to learn and remove various noise sources from images captured by the Lunar Reconnaissance Orbiter. The result was a groundbreaking ML algorithm named HORUS (Hyper-Effective Noise Removal U-Net Software), which enhances orbital imagery of PSRs and allowed scientists and engineers to identify features like craters and boulders in shadowed regions for the first time. HORUS was subsequently applied to help identify potential hazards and plan routes at the Moon's South Pole to enable NASA's VIPER (Volatiles Investigating Polar Exploration Rover) mission to one day achieve the greatest lunar science.

Today, Ignacio is the AI Research Lead and a member of the Intelligent Systems Division (ISD) at NASA’s Ames Research Center in California's Silicon Valley, specializing in AI and robotics to safely and efficiently increase autonomy for space missions. He has been recognized for his career contributions, receiving the NASA Ames Honor Award for excellence in the Contractor Employee category.

NASA COLLABORATION ON AI/ML

Today, KBR is at the forefront of AI and ML development through its collaborations with NASA and other U.S. government agencies— with López-Francos as the tip of our spear. He is currently part of ISRD-3, one of KBR’s flagship AI/ML contracts, which provides NASA Ames’ Intelligent Systems Division (ISD) with technical expertise in scientific research, engineering, and technology development across multiple NASA missions and projects, including VIPER and Astrobee. Since joining ISD in late 2020, Ignacio has contributed to diverse areas such as vision-based robotic navigation, simulation-to-reality, 3D reconstruction, edge computing, neuromorphic computing, anomaly response, and geospatial mapping.

Visual Spatial Intelligence

Since 2022, Ignacio has been co-leading the Neural Radiance Methods (NRM) research group focused on the development of volumetric reconstruction algorithms using both classical computer vision methods as well as novel AI-based methods such as Neural Radiance Fields (NeRFs) to enhance operations and maximize the science in extreme lighting conditions, such as those in the Moon’s poles. NeRFs encode spatial and appearance information and enable high-quality 3D reconstructions and novel view synthesis, pushing the boundaries of visual understanding and rendering.

The NRM group’s projects span multiple phases of lunar exploration, from building digital elevation maps (DEMs) of permanently shadowed regions to guide mission planning and operations, to generating 3D models of lunar regolith samples using multi-view, multi-spectral imagery and be able to infer both physical and chemical properties of the terrain.

His group aims to develop what he calls "visual spatial intelligence and computing", that means not only developing algorithms that enable reconstructing 3D environments and novel view synthesis but also leveraging the latest large language models (LLMs) to extract semantic meaning to the reconstructed scene —and then running these models efficiently on the edge to give space robotic systems real-time situational awareness.

"Picture a near-future scenario in which both mission operators and scientists wear augmented reality glasses to visualize real-time images from the rover’s navigation cameras, monitor telemetry, and interact with a 3D-rendered environment. They could instantly evaluate distances and visualize heat maps pinpointing high-priority targets for sample collection."

Earth Independent Operations

NASA is leading a long-term campaign of human exploration, science, and discovery that begins with the Moon and extends all the way to Mars, as detailed in the Moon-To-Mars Architecture.

A key initiative in this effort is Earth Independent Operations (EIO), funded by the Mars Campaign Office under the Exploration Systems Development Mission Directorate (ESDMD)which oversees crewed deep space exploration. EIO is a multi-year, multi-center project focused on developing integrated technologies that enable crews on Mars to handle problems autonomously, without real-time ground support. Ignacio contributes to the project’s data integration and anomaly response, noting:

“Current missions to the Internation Space Station have ~200-300 people on deck at Mission Control, in constant communication with the astronauts. On Mars, if any issues arise, support from the ground will be delayed by about an hour—20 minutes each way for transmission, plus 15 minutes on the ground to investigate. To support future crewed missions to Mars and other deep space destinations, we have to build the capabilities we currently rely on from Earth.”

This is an area that demands high levels of automation to support future crews, and AI will be increasingly central to enabling such autonomy in everything, from spacecraft diagnosis and resource allocation, to enabling robotic systems to stage cargo and assets before crew arrival.

AI Strategy

Federal agencies are increasingly leveraging AI to enhance public services, bolster national security, and drive scientific progress. In FY2024, the U.S. federal budget for AI R&D stood at $2 billion, a figure expected to rise in FY2025. Following new White House directives—such as Executive Order 14110 on “Safe, Secure, and Trustworthy Development and Use of Artificial Intelligence”—NASA recently announced the creation of the Office of the Chief AI Officer (OCAIO), tasked with responsibly and strategically integrating AI across the agency’s diverse missions and research domains.

Last year, Ignacio was selected to serve as an advisor to the newly formed OCAIO. In this capacity, he counsels NASA’s Chief AI Officer on policy, governance, strategy, talent, and research. The role leverages his substantial experience as an AI practitioner, his Silicon Valley connections, and his unique industry background, while also giving him a platform to advocate for broader investments in AI tools, talent, and fundamental research. He views these commitments as critical for expanding NASA’s capabilities and meeting the grand challenges of space exploration.

Lopez-Francos believes that investing in AI R&D is existential for NASA. One area he sees as poised for transformation is AI agents—autonomous software programs that perceive their environment, make decisions, and execute actions to achieve a specific goal. By bringing these agents into the physical realm—an emerging area known as Embodied AI—they can sense, learn, and act in the physical world:

“It’s not hard to imagine a future where robotic systems support the crew members during missions, taking on tedious or hazardous tasks such as placing equipment, assembling infrastructure, or performing maintenance—freeing astronauts to focus on science and exploration. Today, mining companies already use autonomous haulers and drilling robots, and automotive manufacturers are already integrating humanoid robots on assembly lines.”

AI agents could also coordinate multiple robots, each with complementary capabilities, to collaborate on long-horizon goals—such as exploring lava tubes or other structures that might serve one day as future off world habitats.

Thanks to its location in Silicon Valley, NASA Ames Research Center is uniquely positioned to tap into the region’s top AI labs across academia, Big Tech, and startups. Lopez-Francos and his team already maintain collaborations with UC Berkeley, Stanford, and NVIDIA, among others, and aim to broaden their research partnerships.

As federal agencies face calls to increase operational efficiency, Lopez-Francos hopes to guide NASA in safely adopting AI for day-to-day tasks and draw on his private-sector experience to help further streamline the agency’s workflows.

“NASA needs to harness the latest AI breakthroughs and restore Ames to the forefront of the agency’s AI R&D in order to narrow the gap between Big Tech and government labs.”

After all, Ames not only stands as NASA’s most prolific center for AI research in terms of publications, but it also boasts a decades-long history in AI under visionaries like Dr. Henry Lum or Dr. Peter Norvig. In the 1980s, Ames pioneered AI applications such as autonomous planning and real-time scheduling for mission-critical operations, establishing itself as one of the nation’s most respected AI research groups.

Lopez-Francos believes that stewardship of one of this century’s most revolutionary technologies should not fall solely to Big Tech. To that end, he hopes agencies like NASA research and develop advanced AI in a manner that is both safe and impactful—ultimately ushering in a new golden era of space exploration.