Parsons supports defense and R&D programs across AI and national security. My work focuses on large-scale language model infrastructure and retrieval systems for classified and unstructured government data.

• Build and deploy distributed LLM pipelines for DoD research environments using Python, gRPC, and containerized workflows.

• Develop data extraction and retrieval systems that transform unstructured text into structured, searchable formats.

• Optimize RAG and model-tuning pipelines to improve performance, reliability, and clarity for high-security applications.

• Work across multi-node compute and cloud environments to streamline access and inference at scale. • Supported end-to-end data collection and validation workflows for large-scale systems, ensuring data integrity, traceability, and repeatability across testing runs.

Languages and Tools Utilized: Python, C++, React, gRPC, CUDA, Singularity, RAG Pipelines, Multi-node HPC Systems

Contributed to DoD AI infrastructure development and benchmarking on national supercomputing clusters. Focused on optimizing large language model inference and distributed compute efficiency across diverse architectures.

• Supported dataset organization, labeling, and versioning to maintain consistency across repeated experiments and benchmarking runs.

• Adapted LLM inference to CPU-centric high-performance computing environments using frameworks like GGML and Llama.cpp.

• Benchmarked and tuned LLM performance on supercomputers Chessie (ARL) and Narwhal (Navy DSRC), reducing processing time by up to 40%.

• Tested and profiled model performance across Apple Silicon and DoD HPC clusters to identify optimal CPU/GPU configurations.

• Developed containerized pipelines using Singularity, integrating GitLab, Jupyter, and VSCode within Linux VMs for reproducible experiments.

• Designed multi-node setups leveraging gRPC for distributed communication and analyzed the efficiency of memory-sharing techniques under load.

Languages and Tools Utilized: Python, C++, gRPC, Llama.cpp, GGML, Singularity, GitLab, HPC Systems (Chessie & Narwhal)

Contributed to the development and optimization of software for connected vehicle infrastructure systems integrating real-time data and sensor communication.

• Collaborated with senior engineers to design, implement, and test production-grade software modules in Python and C++.

• Conducted structured hardware–software testing under varied operating conditions, documenting system behavior and collected data for validation.

• Worked across front-end (HTML, React prototypes) and back-end (Python APIs) development to improve application performance and reliability.

• Developed internal tools and visualization dashboards for data interpretation and system debugging.

• Created user-facing prototypes and interface layouts using Figma and Adobe Suite to enhance clarity and user experience.

• Supported testing and firmware integration for the V2I (Vehicle-to-Infrastructure) communication platform.

Languages and Tools Utilized: Python, C++, React, HTML, Figma, Adobe Suite, REST APIs, Connected Vehicle Systems

Conducted interdisciplinary research at the intersection of bioengineering, data science, and environmental modeling. Focused on designing growth experiments and automating data workflows for improved reproducibility.

• Analyzed and visualized multi-sensor datasets using Python and C++ to model environmental and biological responses.

• Built and maintained web dashboards for the Science Department to track and display live research metrics.

• Performed cost and resource analyses for faculty projects to optimize budget allocation and experiment efficiency.

• Supported cross-department collaborations between the Environmental Science and Biology Divisions.

Languages and Tools Utilized: Python, C++, React, Next.js, MySQL, Pandas, Scientific Visualization Tools