James Diffenderfer Email & Phone Number

Livermore, California, United States

• Researching out-of-distribution and adversarial robustness, dynamic neural network adaptation for robustness (to out-of-distribution data) and compression, and catastrophic forgetting; Published at NeurIPS 2022
• Designed neural network pruning and quantization experiments yielding 60× memory and 4× latency reduction on edge device (NVIDIA Jetson) while maintaining 95% of test performance compared to baseline
• Proposed LDRD exploratory research project on gradient-free deep learning and secured funding for team of 4 researchers for 3 years (as Principal Investigator)
• Managing project with team of 4 LLNL research staff, 2 university subcontracts, and 2 summer interns
• Awarded LLNL Deputy Director Science and Technology (DDS&T) Publication Excellence Award

Livermore, California, United States

• Researched and developed neural network pruning and binarization methods and their effects on accuracy and out-of-distribution robustness; Published at ICLR 2021 and NeurIPS 2021
• Investigated theoretical confidence in approximate computing methods through the use of error bound analysis and uncertainty quantification techniques; Implemented strategy on GPUs using CUDA
• Proposed LDRD feasibility study research project on training scientific ML models and secured $150K in funding
• Mentored LLNL Data Science Summer Institute (DSSI) and Approximate Computing summer interns
• Awarded LLNL DDS&T Publication Excellence Award and Center for Applied and Scientific Computing SPOT Award

Livermore, CA

• Investigating efficacy of methods for reducing bandwidth and storage costs in parallelizable algorithms being implemented on high-performance computing clusters
• Collaborated on performing stability analysis for ZFP compression algorithm and for use of ZFP compression inline with parallelizable iterative methods which resulted in 2 research papers
• Communicated results of analysis at the SIAM Conference on Computational Science and Engineering (CSE19) in Spokane, WA

Livermore, CA

• Supported efforts of Variable Precision Computing group on 2 research projects and acquired experience working with high-performance computing systems
• Conceptualized method for analyzing the stability of the ZFP compression/decompression algorithm
• Implemented and tested mixed precision versions of the fixed point iterative solver KINSOL with Anderson acceleration in the SUNDIALS C++ software package that resulted in a reduction in the number of iterations in 25%.

Gainesville, Florida Area

• Researched and developed a nonlinear program solver, NPASA, which resulted in 2 research papers and currently developing C implementation of nonlinear program solver NPASA
• Contributed to the source code and data structure development for the C implementation of the polyhedral constrained optimization problem solver PASA
• Developed makefiles and implemented Python and MATLAB interfaces for the optimization software suite SuiteOPT

Gainesville, Florida Area

• Assisted with instruction and grading for 60 - 120 students 1 semester a year in Calculus 1, Calculus 3, or Differential Equations and received overall instructor rating of 4.76 / 5 from students
• Introduced online student surveys, analyzed responses, and tailored instruction to the needs of each section
• Administered examinations for 60 - 120 students, proofed examinations, and designed weekly assignments

Statesboro, GA

• Instructed and developed course materials for 3 to 4 Calculus I, Survey of Calculus, Precalculus, Trigonometry, or College Algebra courses per semester
• Personalized teaching techniques during office hours to accommodate learning styles of over 50 individuals
• Volunteered on committee to organize and host the Annual Invitational Georgia Southern Mathematics Tournament and facilitated the grading and ranking of 800 - 1000 examinations

Statesboro, GA

• Developed 16 hours of instructional materials for course reviewing math on the Graduate Records Examination (GRE)
• Taught 6 separate month long courses to prepare students for the mathematics section of the GRE

Statesboro, GA

Statesboro, GA

• Collaborated on preparation and organization of bulk mail projects with up to 3,000 letters and compiled visiting student information packets for the Office of Admissions
• Communicated details about Georgia Southern University to over 30 prospective or current students in person and on the phone each week

Statesboro, GA

• Presented information about Georgia Southern University to groups of up to 300 incoming students and parents
• Facilitated 14 different break out sessions for groups of 20 - 30 incoming students
• Supervised 10 enrollment sessions of over 100 students and provided academic counseling to over 50 students

Doctor Of Philosophy - Phd, Applied Mathematics - Optimization, 4.0

Activities and Societies: Society for Industrial and Applied Mathematics (SIAM), Graduate Student Mentor

Master Of Science - Ms, Computer Science, 4.0

Master Of Science - Ms, Mathematics, 4.0

Bachelor Of Science - Bs, Mathematics, 3.79