My name is Arup Sarker, and I am working as a research assistant at the UVA biocomplexity Institute. My research focuses on high-performance data engineering, distributed computing, and deep learning—areas that are critical to the scale at which any system operates its personalization and content delivery systems.

Since joining my research group, I have specialized in optimizing the intersection of data processing and model execution. My work on the Deep Cylon project—building upon the Apache Arrow framework—focuses on maximizing AI algorithm performance through vectorization and efficient GPU utilization. This expertise in high-performance computing (HPC) is directly applicable to system's need for scalable, low-latency machine learning pipelines that power everything from recommendation engines to content encoding.

My portfolio includes several high-impact projects that demonstrate my ability to bridge the gap between theoretical research and scalable engineering:

• Agentic DRC: Developed a unified distributed runtime for LLM and RAG pipelines to eliminate serialization bottlenecks and ensure reproducible, high-throughput execution.

• Deep RC & Radical-Cylon: Engineered heterogeneous runtime systems that integrate data engineering with PyTorch and TensorFlow, reducing execution latency by up to 75.9 seconds and scaling to datasets of 3.5 billion rows.

• Industry Foundation: My research is backed by a patent in image processing from my tenure at Samsung Research and a proven track record of mentorship, currently supervising eight students on the Deep Cylon project.

Before UVA, I worked as a Staff Engineer at Samsung R&D Institute Bangladesh Ltd. in September 2011. It has been a long journey with Samsung, and I have had multiple opportunities to work with different cutting-edge technologies that lead the market. Contributions to Android platform development, especially on Audio/Video, Frameworks, Network protocols to transmit data in mission-critical services and Gear360, and so many others. From the beginning at Samsung, all of my work is aligned with designing and developing embedded and distributed solutions that need to transmit data over networks. Even when I was a developer of a feature phone in 2011, I took care of system modules for transmitting calls and messages.

In addition, I was also responsible for developing Mission Critical Services for first responders and different security agencies. This project will replace the legacy systems of security agencies and ensure public safety in better ways by handling different emergencies. This large-scale distributed system will ensure all data communication, including Audio and Video, with the lowest possible latency. Every single second is very important for the first responder. So ensuring all KPIs for data quality is a major challenging job. We have completed the first milestone and delivered the product.