Computer Science student focused on systems, HPC, and DSP

Byte-oriented file distribution and reconstruction work focused on deterministic chunk layout, fault-tolerant recovery, and practical security constraints.

• Python
• File I/O
• Systems Design
• Cryptography

• Designed around byte-level chunking and scatter/gather reconstruction so payloads can be split, stored, and recovered with predictable structure instead of opaque blobs.
• Focused on efficient split and reassembly paths with attention to memory overhead, streaming behavior, and avoiding unnecessary data copies for larger files.
• Integrated encrypted metadata and fault-tolerant distribution decisions to balance recoverability with zero-knowledge style protection goals.

Distributed linear algebra engine in C covering matrix multiplication, REF, RREF, and LU decomposition with MPI-based work partitioning and benchmark-driven tuning.

• C
• MPI
• OpenMP
• HPC

• Implemented distributed kernels for dense matrix operations with explicit rank coordination, data partitioning, and synchronization across MPI processes.
• Built runtime benchmarks across matrix sizes and process counts to measure speedup, scaling behavior, and the point where communication overhead begins to dominate.
• Tuned memory layout and communication strategy to keep hot loops compute-dense and reduce unnecessary coordination in performance-critical paths.

Interactive audio processing pipeline for experimenting with filters, spectral transforms, and inspectable signal stages from ingestion through processed output.

• Python
• Flask
• FFT
• Audio Processing

• Implemented filtering and FFT-based analysis in Python, exposing waveform and frequency-domain views through an interactive frontend for fast iteration and validation.
• Built an end-to-end pipeline that accepts audio input, applies configurable processing, and returns transformed output rather than limiting the project to isolated scripts.
• Structured the tool so intermediate signal behavior is visible, making it practical for comparing filter choices and debugging spectral changes instead of treating DSP as a black box.