The University of Texas at Dallas
Bachelor of Science in Software Engineering
Jan 2023 - May 2025Software Engineer & Data Science Graduate Student
Logan Margabandu
An MS Data Science student at UTD, blending machine learning with full-stack web development to craft future-ready applications. With a toolkit spanning NextJS, FastAPI, and Google Cloud, I refine my skills through intense hackathons. From civic tech platforms empowering nonprofits to AI-driven healthcare solutions, I'm passionate about building applications that make a tangible difference in society. Explore with me the symphony of code and creativity!
Currently exploring
- Differential privacy-preserving federated learning architectures for distributed financial systems
- Multi-agent reinforcement learning with hierarchical policy decomposition and emergent cooperation
- Quantized transformer inference with dynamic KV-cache optimization for edge deployment at scale
Education
Academic foundations that ground my work in data science rigor and software engineering craft.
The University of Texas at Dallas
Master of Science in Data Science and Statistics
Aug 2025 - Expected May 2027Experience
Blending high-performance computing, applied ML research, and full-stack delivery.
High Performance Computing Software Engineer Intern
Federal Reserve Bank of Dallas
Oct 2025 - Present | Dallas, TX- Research deep learning and machine learning applications for macroeconomic models, optimizing legacy Fortran and C++ pipelines for modern HPC environments.
- Analyze and enhance linear and non-linear financial models to improve computational efficiency and accuracy for banking operations.
Machine Learning Research Intern
National Science Foundation
May 2025 - Aug 2025 | Dallas, TX- Advanced an interactive DoE Oracle app that integrates design of experiments techniques with LLMs, achieving 90%+ automated test accuracy and cutting validation time by 60%.
- Designed over 50 experiments across seven combinatorial strategies, producing insights that lower QA costs in ML systems by up to 40%.
- Authored a peer-reviewed IEEE conference paper that informs emerging ML testing standards for safety-critical industries.
Software Engineer Intern
Fellows Consulting Group
Jan 2025 - May 2025 | Dallas, TX- Delivered a containerized full-stack application that reduced deployment time by 40% while maintaining 99.9% uptime for client-facing workloads.
- Designed and fine-tuned a DoE Oracle LLM (Llama 3.2) to cut manual test-case validation effort by 60%, redirecting consultant time to higher-value work.
- Improved model accuracy by 35% across 10K+ scenarios through multi-framework integration, boosting product reliability and reducing production risk.
Information Technology Support Specialist
The University of Texas at Dallas Office of Technology and Innovation
July 2024 - May 2025 | Dallas, TX- Delivered technical support to 1,000+ students across two of UT Dallas's largest labs (30,000+ student body), achieving a 98% first-contact resolution rate and ensuring minimal disruption to academic operations.
- Streamlined IT asset management by performing periodic inventories of 200+ departmental devices (computers, lab equipment, peripherals), improving tracking accuracy and reducing missing-equipment incidents by 15%.
- Improved lab efficiency and student experience by ensuring labs were organized, stocked with supplies, and fully operational, cutting downtime due to missing resources by 20%.
- Enhanced reliability of IT operations by identifying and reporting hardware/software failures, troubleshooting networking issues, and escalating complex problems, resolving issues 30% faster.
Projects
Selected builds spanning AI reasoning, civic tech, and research-driven automation.
01
🏆 3rd Place - HackUTA
Nomi.ai
An AI reasoning system delivering biomarker insights and voice-first care guidance for patients. Achieved 90% accuracy in symptom insights by fusing multi-model reasoning with medical knowledge graphs and integrated natural voice interaction for real-time analysis.
- Node.js
- Claude 3.5
- Gemini API
- ElevenLabs
02
AVANCE Dallas Grant Finder
A bilingual platform that centralizes local grant data and streamlines nonprofit application workflows. Accelerates grant application timelines by 40% with guided tracking, aggregating 500+ local grants annually through automated scraping and deduplication.
- React
- FastAPI
- Scrapy
- PostgreSQL
03
🏆 IEEE Conference Publication
A Design of Experiments Oracle LLM
Research-grade experimentation harness that blends design of experiments with LLM evaluation to boost QA rigor. Runs 50+ structured experiments to deliver 60% faster validation loops and lifts model accuracy by 35% across 10K+ test scenarios.
- Python
- Llama 3.2
- LangChain
- LLM QA
Skills
A toolkit shaped by research, production systems, and community leadership.
Languages
Machine Learning
Frameworks & Cloud
Contact
Have an opportunity, collaboration idea, or challenge? Let's talk.
Drop a message
I'm open to data science, AI research, and software engineering roles, along with speaking and hackathon mentorship opportunities.