Internships

Duane Edgington

Automated classification of deep-sea imagery: MBARI has a rich collection of underwater video and photographs, much of which has been professionally analyzed and curated. We are exploring state-of-the-art automated classification and analysis techniques. This intern will join us in this exploration, testing selected techniques against collections of underwater videos or images to detect and classify organisms of interest to MBARI scientists. One area we are exploring is weakly supervised methods. A background in computer science is required; coursework or experience in machine learning and computer vision would be an ideal background.

Gene Massion

Autonomous Coastal Profiling Float: We have a broad spectrum of potential projects spanning a range of disciplines suitable for a summer intern. An overview of the project can be found at https://www.mbari.org/coastal-profiling-float/.  We are looking for an intern with some experience and a strong interest in one or more of the following topics and a particular interest in developing technology for oceanographic research applications.

1) Design and development of embedded microcontroller based systems.  The ST Microelectronics STM32 family is of particular interest.
2) Design and development of embedded C/C++ software.  Training and/or experience in rigorous software testing methodologies is of particular interest.
3) Automated test systems and web based applications using LabView.
4) Mechanical design of robotic oceanographic research equipment using Solidworks CAD tools

Steve Haddock

Biodiversity and Biooptics: Steve Haddock’s lab aims to characterize and monitor the diversity and behavior of gelatinous plankton (jellyfish and their kin) in the deep sea and open ocean. As an internship project, you will choose what interests you from a related set of topics. These include DNA metabarcoding of plankton samples; biochemistry of bioluminescence; historical time-series analysis of deep-sea video data; generating interactive taxonomic keys; studying how comb jellies (ctenophores) function under high pressure.

Kakani Katija and Joost Daniels

Bioinspired design: In order to fully explore our ocean and discover the life that lives there, we need to scale up our observational capabilities both in time and space. The ocean represents the largest habitable ecosystem on our planet, yet less than 5% of that volume has been explored, and nearly 50% of marine life are yet to be described. To close this gap, scientists are increasingly leaning on underwater imaging systems and robotics to enhance their observational capacity. The Bioinspiration Lab fuses both imaging and robotics, leveraging computer vision and data science to create workflows, data pipelines, and hardware/software tools to expand our understanding of the ocean and its inhabitants in a time of great change. Interns will have the opportunity to: train and deploy machine learning models on ocean visual data in real world settings (see FathomNet Database); conduct human-computer interaction experiments via video gaming (see FathomVerse); or deploy state-of-the-art algorithms on-device for underwater vehicle control at sea or in simulation. Applicants with experience using Python, computer vision, machine learning, web development, and/or embedded systems are preferred.

Jim Barry and Steve Litvin

Seamounts: Research on the biology and ecology of seamount ecosystems, including coral and sponge communities at Sur Ridge off Central California and octopus breeding colonies at hydrothermal warm springs along the foothills of Davidson Seamount are a central focus of research in our lab. Our research ranges from linkages between ocean conditions (e.g., currents, oxygen, temperature, pH, carbon flux) and coral distribution and conditions, to the biology of breeding octopuses. We use various platforms (ROVs, AUVs, oceanographic moorings) and sensors (imaging systems, current meters, chemical sensors) to observe and measure these ecosystems. Opportunities in our lab for a summer internship would fall within this spectrum of research, including analysis of video imagery in relation to high resolution mapping data, the association of corals with current patterns across Sur Ridge, or the biology of animals inhabiting warm springs at the base of Davidson Seamount.

Francisco Chavez

eDNA, otters, and kelp:

Estuaries are regions of enhanced biological dynamics and home to threatened species like sea otters. This project will explore the use of eDNA to measure the effects of Southern Sea Otter predation on estuarine communities. The intern will: Determine the species/taxonomic groups to target. 2) Determine the optimal eDNA methods (qPCR/metabarcoding). 3) Explore the spatial/temporal sampling required to acquire a holistic view of the biodiversity of estuaries and how they are changing. 4) Explore the spatial/temporal sampling required to measure the effects of Southern Sea Otter predation on estuarine communities.

Ocean optics: Our dynamic team seeks an intern interested in linking ocean optics with ocean ecosystem information gleaned from environmental DNA (eDNA). The primary focus will be integrating decade long time series of hyperspectral and eDNA data and developing relations between the color of the ocean and organisms ranging from microscopic plankton to whales. The intern will also have the opportunity to participate in the field collection of optical data, eDNA samples, and the processing of the optical data and eDNA samples onshore. 

Biodiversity: Understanding temporal changes and spatial shifts in species assemblage is key to understanding ocean health and developing management actions. Traditional biodiversity surveys, however, are costly in both time and resources and as a result do not provide the needed coverage in space and time. One of the focuses of our lab is the use of environmental DNA (eDNA) as a tool to track changes in biodiversity. eDNA can detect the presence of a wide variety of organisms from microbes to whales with just a liter of sea water and has the potential to be automated. As part of a Marine Biodiversity Observation Network (MBON) projects eDNA samples have been collected along the California coastline from a variety of environments including kelp forests, seamounts, deep sea, and coastal waters. We seek an intern to explore biodiversity shifts from these datasets. The interns will also have the opportunity to participate in the collection of samples at sea, processing of samples in the molecular lab, and developing/optimizing new assays/primers.

George Matsumoto

GO-BGC Adopt-a-Float: This intern would be responsible for helping to coordinate the Adopt-A-Float program and assist with outreach to educators and other interested partners. This intern would be sponsored by the National Science Foundation as part of the Global Ocean Biogeochemistry Array program and in association with the NSF funded Southern Ocean Carbon and Climate Observations and Modeling (SOCCOM) project.

Akshay Hinduja and Giancarlo Troni

Multi-modal sensing:
“While advances in sonar and imaging technologies have improved our understanding of the seafloor, integrating these data streams presents new opportunities for enhancing underwater surveying systems. This project aims to develop innovative methods for underwater vehicle navigation and mapping, focusing on integrating sonar data with visual information from cameras. The primary objective is to create a novel system for registering camera imagery from smaller, maneuverable autonomous underwater vehicles (AUVs), to sonar-generated sea-floor maps made by surface vessels. This project draws parallels to the birds-eye view (BEV) problem for autonomous ground vehicles in GPS denied environments, adapting those principles to the challenges of underwater mapping and localization.
Candidates should be comfortable with programming in Python or C++. Previous experience with computer vision and machine learning projects is desirable.”

Giancarlo Troni and Sebastian Rodriguez

Scalable Marine Robotics: Ocean exploration has been widely developed thanks to marine robotics, whose platforms are currently being used on several applications, such as accurately mapping the seafloor in high-resolution and continuously tracking animals in midwater. However, these platforms are not scalable, many are still too expensive to build and operate, and access to scientists, and therefore ocean exploration and discovery, is limited due to underwater vehicle navigation among others. Potential intern projects will use current MBARI’s robotics platforms to enable scalable marine robotics navigation in complex terrain. Efforts include sensor calibration and alignment of sensor data and visual-inertial navigation based on simultaneous localization and mapping (SLAM) framework. The work will combine elements of estimation, computer vision, software development and data analysis. Candidates should have basic competence in C/C++ and Python programming. Experience with robotics will be advantageous.

Cassandra Burrier

Science Communication: This project is ideal for an intern with experience translating science and technology concepts into written and visual content for the general public. The intern will work with MBARI staff, scientists, and engineers to develop social media stories about MBARI research. Responsibilities will include creating content from MBARI’s image and video archive and producing compelling stories about our research for our website and social media outlets. We are seeking a candidate with a background in outreach and communication, preferably with interests in science and technology. Strong writing skills and social media content creation skills are required. Experience in video editing, other visual content creation (i.e., animations, infographics, illustrations), and photography is desired. Applications should include writing and multi-media content samples in addition to the other required materials.

Magdalena Carranza

Storm impacts on upper-ocean dynamics and biogeochemistry in the Southern Ocean:Biogeochemical (BGC) Argo floats deployed by the Southern Ocean Carbon and Climate Observations and Modeling (SOCCOM) project have revealed strong CO₂ outgassing in the wintertime, challenging our view of the Southern Ocean as a strong carbon sink. Storms play a major role in driving air-sea CO₂ outgassing, both in ocean models and SOCCOM float observations but the magnitude and phasing of the ocean response to storms differs between models and observations. I am looking for a summer intern willing to explore SOCCOM float data during the passage of the strongest storm recorded in the Southern Hemisphere, traversing the southeast Pacific in October 2022. The goal is to gain mechanistic understanding of bio-physical interaction processes that could lead to air-sea CO₂ exchange in response to such an extreme event. Basic programming skills and/or strong willingness to acquire those skills is required. As a summer intern, you will have the opportunity to learn about biogeochemical sensing capabilities on Argo floats, how to access and analyze BGC Argo data.

John Ryan

Soundscape: Marine animals use sound in essential life activities including communicating, socializing, foraging, navigating and reproducing.  Recording and analyzing sound in the ocean thus provides a profound window into their lives.  MBARI’s sound recordings from the heart of Monterey Bay National Marine Sanctuary have expanded our knowledge of regional biodiversity and revealed complex and beautiful dimensions of animal behavioral ecology.  This project will offer the intern opportunities to wield a variety of advanced analysis methods that are needed to detect and classify specific sounds produced by different species, and to synthesize those results with understanding of the ecosystem in which they live.  Mentors include scientists and software engineers who are actively collaborating at this exciting frontier.  

Lonny Lundsten

MBARI machine learning model performance metrics workflow: MBARI’s video lab has developed numerous object detection machine learning models for automated analysis of deep-sea video footage. In addition, modern data-centric deep learning approaches have shown promise in automatically detecting and correcting label issues in order to improve the quality of training data. Unfortunately, off-the-shelf performance metrics have fallen short when run on our unique video and image datasets. We seek an intern with expertise in data science, machine learning, computer vision, and statistics to help us develop a custom workflow and analysis package for analyzing and reporting on dataset quality and machine learning model performance, tailored to our unique data and analysis workflow. Ideally these metrics would 1) be used for reporting on model performance when evaluated against a human-annotated, gold-standard benchmark, 2) help detect likely issues in the training data, and 3) provide insight as to how model performance changes with updates to the training set (i.e., more training data is added or selective removal of poor-quality training data). We envision a documented analytics workflow and codebase (e.g., Python notebook) as the final product. Candidates should have a diverse technical skillset with expertise in data science and statistics while also being proficient in Python or other computer programming languages.