Resume

NATHAN KONG

*******@***.*** 408-***-**** San Francisco, CA

Education

Penn State University (University Park) – College of Engineering State College, PA

• B.S. in Computer Engineering GPA: 3.5/4.0

• Coursework: Data Structures & Algorithms, Computer Architecture, Systems Programming, Digital Circuits

• Clubs: Nittany AI, Stanford AI Summer Camp, Claude Builder Club, Investment Club Work Experience

Hexagon – Software Engineer Intern May - August 2025

• Created an agentic RAG framework for context-aware recommendations and led the development of company’s first fully automated LLM workflow, now core to its monetized framework

• Gained experience in building Python / LLM workflows, translating business use cases into platform features

• Enriched structured data and context by integrating Azure OpenAI plugins, graph-based agents, and automated pipelines

Research

Penn State Autonomous Vehicle Team – Perception/Computer Vision Spring 2026 (Incoming)

• Accepted to a competitive engineering team developing perception systems for autonomous vehicle navigation

• Emerging opportunity to develop computer vision and ML algorithms to process sensor data and interpret vehicle surroundings for real-time object detection and environment mapping

• Seeking hands-on experience to build messaging pipelines using ROS2 to integrate perception outputs with Guidance, Navigation, and Controls (GNC) systems for autonomous decision-making FlashAttention Scaling and Kernel Optimization Summer 2025

• Authored independent research paper analyzing FlashAttention’s role in enabling 4 longer sequences (up to 131k tokens) on GPT-NeoX-20B across 8 A100 GPUs

• Developed and tested fused attention kernels in Triton, applying persistent kernel execution, warp specialization, and block tiling to improve memory locality and throughput

• Benchmarked experimental kernels against FlashAttention-2, highlighting performance tradeoffs in long-context inference workloads

Projects

Elevator Controller: Python to Vivado HLS Conversion Fall 2025

• Explored conversion of SCAN elevator algorithm from Python to Vivado HLS, analyzing differences between software and hardware implementations

• Developed Python based dynamic heap request management using bidirectional queues and state caching

• Uncovered HLS limitations with the use of Heap operations, later replaced with FIFO to accommodate FPGA synthesis requirements

Skills & Certifications

• Programming/AI: Python, Java, C, SQL, Triton, CUDA, PyTorch, Pandas, NumPy, Scikit-learn, XGBoost, LightGBM, VADER, TextBlob, ChromaDB, Ollama

• Cloud/Systems: Docker, Azure OpenAI, Azure Container Instances, REST APIs, FastAPI, ESPN API, Git

• Hardware/Logic Design: RTL Design, AMD Vivado, Vivado HLS/Pragmas, FPGA, Fixed-point arithmetic

• Certifications: Hugging Face MCP, AI Agents (Vanderbilt), PSU ML Bootcamp, Coursera (Java, Python) 1

Contact this candidate