ExperienceSoftware Engineering InternJune 2026 – PresentRelegance Inc.College Station, TX Engineered real-time WITS/viscometer ingestion pipeline on Azure, streaming multi-edge sensor data into PostgreSQL Built Model Evaluation Workbench with Azure Kata Containers, supporting 4 formats and RMSE/MAE/R2 metrics Optimized FFT signal-processing pipeline, boosting spectral resolution 100x and cutting cold-start latency 4xUndergraduate ResearcherJanuary 2025 – PresentTexas A&M Internet Research LabCollege Station, TX Developed a VQ-VAE model using PyTorch, trained on CIFAR-10 with 30% improved sample efficiency in RL tasks. Contributed to batch processing pipelines enabling object-oriented modeling across diverse scalable RL environments. Designed a Neural Logic Machine with async execution, processing 10,000+ interactions with optimized throughput.Software EngineerJanuary 2024 – December 2024Texas A&M Soil & Crop SciencesCollege Station, TX Developed CNN CV system with PyTorch on AWS SageMaker, achieving 87% classification accuracy on drone imagery. Built scalable signal pipeline on AWS EC2, analyzing 12-band spectral data across 10K+ cotton plant samples. Optimized feature extraction with PCA and automated error handling, reducing false positives by 23%.Undergraduate ResearcherJanuary 2024 – May 2024Texas A&M University Department of StatisticsCollege Station, TX Built end-to-end ML pipeline with Python, processing large-scale RNA data using scalable batch processing architecture. Developed automated data workflows with PostgreSQL, improving reproducibility by 40% across cohorts. Engineered an aneuploidy variational algorithm processing 6M+ cells for real-time data retrieval.ProjectsKDB+ Tick Data ArchitectureJuly 2026 Engineered end-to-end kdb+ market data platform, processing 10M+ live trade and quote events Built high-performance C++ feed handler with batching and fault recovery, sustaining 1M+ msgs/sec Implemented VWAP, OHLC, and trade-quote as-of joins for institutional market analytics Integrated PyKX research layer, achieving 20x+ faster analytical queries than PandasCapital One – Private Finance DashboardJanuary 2026 Selected as 1 of 80 for Capital One Technology Summit, Winning 1st Place for best project Built full-stack finance app with Next.js and Express.js, processing real-time transactions for 500+ simulated users Coded event-driven architecture with four AI agents, reducing API calls by 90% through smart triggering Implemented data aggregation layer and prompt caching, reducing token usage by 40xRestaurant POS SystemMay 2025 Spearheaded the development of a point-of-sale system using Java and PostgreSQL utilizing Agile and Scrum practices Implemented Model View Controller architecture with Java Swing for responsive UI, reducing screen load times by 40% Designed automated inventory management system tracking 100+ products, triggering alerts at configurable thresholds Wrote a Python data generation script to simulate 39 weeks of business operations, producing 750K+ in sales dataSkillsLanguages: C++, Python, Java, JavaScript/TypeScript, C#, R, SQL, GitFrameworks & Libraries: PyTorch, Pandas, CUDA, NumPy, Scikit-learn, React, FastAPI, Node.js/Express.jsCloud & DevOps: AWS, Azure, GCP, Docker, Kubernetes, Terraform, CloudFormation, CI/CD, GitHub ActionsSystems & Performance: Data Structures & Algorithms, System Design, Concurrency & Multithreading,Distributed Systems, Low-Latency System Design, Networking (TCP/IP, REST), Memory Optimization, Profiling& Benchmarking