Student R&D Opportunities

The Universities Space Research Association (USRA) in collaboration with the NASA Advanced Supercomputing (NAS) Center at NASA's Ames Research Center (see nas.nasa.gov), is seeking student interns to conduct collaborative research and workforce development projects related to high performance computing.

The objectives of the internships will be to identify, develop, and (where appropriate) demonstrate advanced supercomputing capabilities that would have practical value for enhancing NAS Services in a two- to five-year timeframe. NAS Services include high-end computing, storage, networking, and associated capabilities that enable scientists and engineers supporting NASA missions in space exploration, scientific discovery, and aeronautics research.

In your cover letter, please rank three (3) projects listed below from most to least interesting to you and include a personal statement describing why you are interested in these particular projects, what you hope to gain from the internship, and what makes you qualified for this internship. Please provide links to a cover letter, your CV and an unofficial transcript. Provide a link to your unofficial transcript in the last question (i.e. additional materials).

• Performance prediction models
• Benchmarks for new architectures
• Hardware infrastructure, file system technologies and/or application enhancements for optimized I/O
• Programming approaches for accelerated and non-accelerated architectures
• AI/ML algorithms to exploit traditional hardware architectures
• Mixed/variable precision computing and machine learning/AI for scientific computation
• Optimized techniques/tools for visualization, data analytics, artificial intelligence and machine learning
• Hardware enhancement for pathfinding systems
• Efficient math libraries
• Software packaging technology, such as containers
• Algorithms to exploit new hardware enhancements
• Lossy compression for scientific datasets
• Methodologies and tools for application performance optimization
• Frameworks for distributed data depositories
• Science data processing pipeline frameworks
• Frameworks for hybrid cloud/on-prem scientific workflows

The goal of this project is to build a similarity search engine for multivariate time series of commercial flights. The volume of recorded flights data is increasing exponentially and reviewing and labeling such data is largely expensive and impractical and requires significant amounts of effort from subject matter experts. On the other hand, if there is a search engine specifically designed and tuned for flights, one can input a data of interest (which is potentially reviewed and labeled by the experts) and efficiently search through the entire database to find patterns similar to the data of interest.

To be able to do this, we first need to build a low-dimensional embedding of the high- dimensional and heterogeneous flight time series, where the search can be performed efficiently. The project has two main components: (1) a self-supervised approach based on deep neural network to provide a low-dimensional embedding for the flight time series, and (2) a similarity search algorithm that can efficiently search for similar data to the query one in the embedding space. Several metrics such as precision at k can be used to evaluate the performance of the developed model.

This work is in support of the Data Sciences Group at the Intelligent Systems Division at NASA Ames Research Center. The student intern will learn to develop machine learning models for dimensionality reduction and ranking in flight’s operational quality assurance data. The real-world commercial flights data will be provided for the project.

The outcome of this internship is beneficial to both NASA and the student. The developed materials will help advancing the system-wide safety assurance project of NASA, specifically in identifying precursors for flight safety. Moreover, this internship is beneficial to the student in several ways: (1) student will get hands-on experience on implementing state-of-the-art machine learning models on real-world databases, and (2) get involved in research projects from the beginning and learn how to communicate the findings efficiently in research code as well as scientific publication and presentation.

• Programming in Python (Demonstrated ability to write research-level code).
• Data science including data wrangling, cleaning, and visualization (pandas, numpy, matplotlib).
• Implement deep neural networks with Keras, Tensorflow and/or PyTorch.
• Demonstrated ability to work in small teams.
• Excellent oral and written communication skills.
• Should be pursuing a graduate degree related to Computer/Data Science or Engineering.
• Works well in a remote working environment over MS Teams.

Up to 6 interns will be selected for this internship and will be split into 3 groups. Each group will work on one of the following 3 projects. In your cover letter, please rank the 3 projects listed below from most to least interesting to you, why you are interested in this internship, what you hope to gain from the internship, and what makes you qualified for this internship. Please provide links to a cover letter, your CV and an unofficial transcript. Provide a link to your unofficial transcript in the last question (i.e. additional materials). 

• Remove Duplicates using UMIs: Process datasets containing Unique Molecular Identifiers (UMIs), then compare the results if you remove duplicates vs keeping them in.
• In silico rRNA Removal: Select two sets of GeneLab datasets (set 1: datasets where most samples contain high rRNA contamination; set 2: datasets containing very few samples with high rRNA contamination) and re-process them but this time remove rRNA with HTStream - compare the results to those without removing rRNA. Next, compare HTStream rRNA removal (k-mer based) vs. removing rRNAs with an aligner (i.e. STAR) based removal process.
• Single-Cell (and/or Spatial Transcriptomic) RNAseq processing: Help develop and test pipelines for processing single-cell RNA sequencing data and/or Spatial Transcriptomics data.

• Successful completion of San Jose State University's Bioinformatics ll (Biol 23B / CS 123B) course
• Successful completion of the GeneLab RNAseq bootcamp
• Good academic standing

This project is to develop a nominal prototype of an “Autonomous Entity Global Intelligence System” or AEGIS, by incorporating cutting edge Artificial Intelligence (AI) and Machine Learning (ML) technologies to enable a distributed supervisory control framework for Urban Air Mobility (UAM). In this project, the student will apply cutting-edge AI/ML technology to make real-world aviation and coordination among autonomous air vehicles efficient and safe. The intern will deliver working data synthesis, simulation, and visualization modules in the form of software codes.

This position will consist the following tasks:

• Create representative sets of operation scenarios, by synthesizing existing aeronautics and aviation data
• Develop a machine learning model for analyzing and inferring the synthesized data
• Setup a simulation testbed to ingest, engineer, infer the data, and visualize the simulation results
• Implement and evaluate multiple operational modes in different unconstrained and constrained scenarios

Category A

• Desired skills: D3.js, MATLAB, C/C++
• Additional preferred skills: Visual Studio, Unity, Adobe Illustrator
• Experience level: 0-5 years

Category B

• Desired skills: Python, MATLAB, Tensor Flow, CUDA
• Additional preferred skills: KERAS, Open CV, Unity
• Experience level: 0-5 years

Additional general expectations include:

The Air Traffic Management Exploration (ATM-X) project will develop and demonstrate a new service-based air traffic system paradigm. NASA researchers are exploring an initial concept development of Urban Air Transport Disruption Management platform (UATDM). UATDM aims at enabling highly automated commercial services provided by Unmanned Aircraft Systems and Urban Air Mobility in low-altitude airspaces. The selected interns will have the opportunities to perform researches and develop algorithms for UATDM.

Research, poster presentation, numerical models or algorithms, report or conferences/journal publications

• Mobile App development knowledge or experiences in Matlab/C/C++/Java/Optimal controls/machine learning

This work supports the UAS in the National Airspace System (NAS) integration project. The goal is to characterize detect and avoid (DAA) system behavior and identify / extract scenarios of interest in order to further the project's research objectives.

To accomplish this, the work will focus on visualization and analysis of DAA encounters between UAS and manned traffic operating under visual flight rules.

The UTM (Unmanned Aircraft System (UAS) Traffic Management (UTM)) research platform enables the testing of various capabilities ranging from visual line-of-sight operations in rural areas to beyond visual line-of-sight (expanded) operations in urban areas. This platform is comprised of a set of web services that are accessible from a remote server by clients that have been implemented according to the published interface control document (ICD). The ICD defines the communication between the UTM research platform and an operator. It provides the information necessary to develop clients and software applications that interface with the UTM research platform.

The UTM research team has developed a number of clients and applications that displays UTM information and enables interaction with the UTM research platform. The goal is to distribute this client software to as many universities as possible and assist them in communicating with the UTM server at NASA Ames to help enhance the UTM concept development and provide a wider base for research within the academia in this field.

USRA's R&D Student Program supports projects at NASA Ames as part of the NASA Academic Mission Services (NAMS). The goal of these internships is to help advance the milestones of the project and to help complete the student's thesis.

The internships range in technical areas of:

• Aeronautics
• Earth Science
• Data Science
• Machine Learning
• Robotics
• Synthetic Biology
• Biosciences
• Environmental Analytics