Jack Alvarez Email & Phone Number

Fairfax, Virginia, Us

▪ Processed and decisioned a portfolio of several thousand business loans, personally facilitating the disbursement of over $31.4 million dollars in critical funding during the COVID-19 pandemic. Analyzed business cash flow, tax data, credit data, business viability, fraud indicators, and compliance documentation to determine loan eligibility and borrowing capacity. ▪ Programmed a modular codebase in VB.NET to extract, transform, and load application data from the client’s web database into deployment environments for robotic process automata (RPA). Engineered and deployed RPAs to independently execute various complex analyses and business processes. Doubled team production speed and significantly enhanced service quality and reliability. ▪ Designed and implemented a web-based production tracking and timekeeping system with VB.NET. Acted as technical support for a 100-person team and investigated and resolved technical issues. Developed tracking software for technical issues. Produced and delivered technical support documentation and presentations. ▪ Gathered, compiled, and analyzed evidence of financial crimes and fraud. Delivered comprehensive reports to investigators and federal attorneys detailing findings. Developed a natural language processing pipeline which enabled the detection of mass-impersonation cyber threats.

Blacksburg, Va, Us

▪ Successfully developed and delivered a machine learning solution to the black box Algebraic Multigrid (AMG) optimization problem. Conducted novel research at the intersection of machine learning, numerical simulation, and graph theory. Designed and implemented a novel and powerful method to accelerate the performance of MueLu, the AMG solver in Sandia’s popular Trilinos library for scientific computing.▪ Consulted with the client and stakeholders to identify the project scope, goals, and key deliverables. Continuously planned, executed, and evaluated each project phase. Delivered dozens of technical papers, progress reports, data visualizations, and demonstrative presentations to the client, stakeholders, and academic faculty.▪ Collaborated with a cross-functional team of researchers and academic faculty to develop robust and efficient optimization methods in parallel computing environments. Analyzed several hundred pages of academic papers and software documentation to develop an ideal pipeline for the AMG problem.▪ Generated thousands of randomized test cases by automating linear system generation in C++. Fit optimal AMG parameters for each test case with a parallel swarm optimization algorithm in C++. Developed a robust method to reduce arbitrarily large linear systems to a fixed-dimensional kernel estimation.▪ Implemented a neural network in TensorFlow using Python to predict optimal AMG parameters given any solvable linear system. Trained the network to estimate AMG parameters which minimize solver execution time given matrix kernel estimators. Consistently reduced solver iterations by up to 80% and achieved zero-shot convergence against failed baseline tests.

Blacksburg, Va, Us

▪ Evaluated Finite Element solver performance between various preconditioning methods in Conjugate Gradient and other linear system solvers. Tested and benchmarked a suite of preconditioners and solvers in the deal.II and Trilinos finite element libraries under shared and distributed-memory parallelism.▪ Analyzed the effects of preconditioning method, relaxation parameters, drop tolerance, mesh anisotropy, node reordering and adaptive mesh refinement on matrix condition number decay and solver stability.▪ Invented a novel distributed-memory parallel algorithm to solve large linear systems using sparse reordering techniques and algebraic domain decomposition. Implemented a naïve implementation in MATLAB and benchmarked performance against traditional solving methods.

Washington, Dc, Us

▪ Designed and programmed bookkeeping software in VB.NET to generate global financial reports for the Air-to-Ground Communications team. Implemented context-preserving unique identifiers enable precise tracking of all transactions. Automated calculation of important financial metrics and discrepancy tracking. Wrote scripts in VB.NET to extract, transform, validate, and load transaction data from invoices in various document formats to target ledgers.▪ Generated intuitive and informative data visualizations for team leaders for a variety of projects.▪ Wrote and published Standard Operating Procedures, best practices, and support documentation for software design, data management, and numerical error handling. Presented accomplishments to team leaders and both employer and customer executives.