Mithun Karthikeyan
AI & Software Engineer
I build reliable AI systems, research tools, and full-stack products that turn complex information into useful, inspectable experiences.
- 2026
- M.S. Artificial Intelligence
- 4.0
- B.S. Computer Science GPA
- 03
- Peer-reviewed publications
01 / About
Engineering grounded in evidence.
I am an AI and software engineer with experience spanning machine learning, systems programming, scientific computing, cloud infrastructure, and product development.
My work moves between research and production: from peer-reviewed NMR software and computational analysis to RAG pipelines, data-backed applications, and agentic AI experiences. I care about systems that are not only capable, but understandable, testable, and useful to the people relying on them.
Bachelor of Science, Computer Science - Data Science
Graduated with a 4.0 GPA.
Languages
- Python
- C / C++
- Java
- TypeScript
- SQL
- R
AI & data
- PyTorch
- Scikit-learn
- Pandas
- NumPy
- SciPy
- ChromaDB
Product engineering
- React
- Node.js
- FastAPI
- Flask
- AWS
- GitHub Actions
Systems & research
- Linux / Unix
- OpenMP
- CUDA
- HPC
- NMR analysis
- LaTeX
02 / Experience
Teaching, research, and delivery.
Graduate Teaching Assistant
Oregon State University
Corvallis, OregonTaught and mentored across systems, open-source software, and senior capstone courses.
- Supported graduate students building and debugging xv6 kernel features for Operating Systems II.
- Mentored open-source contributors on Git workflows, pull-request review, licensing, modular design, testing, and CI.
- Advised senior capstone teams from requirements through deployment, with an emphasis on architecture, documentation, and delivery.
Research Assistant
Osborn Popp Lab, Oregon State University
Corvallis, OregonBuilt open-source scientific software and contributed to peer-reviewed high-pressure NMR research.
- Developed NMR Plotter, a Python GUI for publication-quality multi-spectrum figures with reusable plotting settings.
- Created a Python tool for simulating and fitting CODEX decay data to extract phospholipid diffusion coefficients.
- Improved research workflows through reliable data processing, error handling, visualization, and reproducible software practices.
- Metallized, 3D-printed radiofrequency coilsJournal of Magnetic Resonance, 2026
- NMR_Plotter: A Python GUI for Generating Figures from NMR SpectraJournal of Open Research Software, 2025
- Measurement of phospholipid lateral diffusion at high pressure by in situ magic-angle spinning NMR spectroscopyCommunications Chemistry, 2025
Web Development Intern
RUSD Solutions
RemoteContributed to e-commerce delivery, WordPress experimentation, and workflow automation.
- Supported the deployment of an e-commerce site and contributed practical design recommendations.
- Integrated and troubleshot an AI Engine plugin in a WordPress test environment.
- Automated Adobe Acrobat form-filling workflows to reduce repetitive manual work.
03 / Projects
Selected systems and tools.

SmartTap
A natural-language evidence system that turns questions about Oregon agricultural and weather data into charts, summaries, previews, and inspectable validation artifacts.

Mocki
A first-place BeaverHacks project, now live as a resume-aware interview practice platform with a rotating AI panel, voice practice, and actionable debriefs.

LLM Agents at the Edge
An HP-sponsored RAG system for Oregon State engineering students, backed by a data pipeline, vector retrieval, and scalable AWS infrastructure.

NMR Plotter
An open-source desktop application for creating publication-quality NMR spectra figures with multi-spectrum layouts and reusable styling controls.

Dog Emotion Detection
A transfer-learning study comparing ResNet-18 and ResNet-50 across fine-tuning and augmentation strategies, reaching 86.2% validation accuracy.

Roll Call Brawl
A 24-hour hackathon project that turned classroom attendance into a competitive, geolocation-validated check-in experience.

CODEX Decay Curve Fitting GUI
An open-source scientific tool for simulating and fitting CODEX decay data used in high-pressure NMR diffusion research.
04 / Contact
Let's build something thoughtful.
I am always interested in meaningful engineering, applied AI, and research collaborations.
mithunkar2004@gmail.com