For engineering teams
Making AI work when the demo is over.
I help teams design, evaluate, and productionize AI systems — particularly agentic workflows, RAG, evaluation infrastructure, and Responsible AI readiness.
Salt Lake City, Utah Engineering & research
Building and studying AI systems for the real world.
Victoria Doyon
AI Systems Engineer · Researcher · MS CS @ U of U
I'm Victoria Doyon — an AI systems engineer and M.S. Computer Science student working across agentic AI, retrieval, evaluation, and production ML systems. My work sits at the intersection of building AI that works in the real world and studying how we can make intelligent systems more reliable, understandable, and capable.
To me, these aren't separate interests. Better research informs better systems, and production systems expose the questions worth researching.
For research
Understanding how we can make AI better.
I'm interested in research around LLM and agent evaluation, retrieval, NLP, explainable and Responsible AI, and the intersection of machine learning and neuroscience.
Focus areas
I work on the parts of AI systems where model behavior meets software engineering: retrieval, agents, evaluation, infrastructure, and the safeguards that make them dependable.
Focus area
Agent Evaluation & Reliability
Testing how agentic systems behave, where they fail, and whether they can be trusted outside controlled demos.
Related researchFocus area
Retrieval & Grounded AI
Designing retrieval systems that give models the right information at the right time — and measuring whether that grounding actually improves the result.
Related researchFocus area
AI Systems Architecture
Taking AI capabilities from prototype to production through scalable architecture, deployment, testing, security, and operational readiness.
Experience
Experience
2026 — Present
University of Utah / SCI Institute
Program Assistant (CHPC)
Experience
2024 — 2026
Microsoft
Software Engineer II
Agentic AI · RAG & AI Search · Responsible AI · AI infrastructure · Production Reliability
Experience
2022 — 2024
Microsoft
Software Engineer
Full-stack systems · Anomaly Detection · Cloud Infrastructure · Security · Operational Engineering
Experience
2021 — 2022
HereAfter AI
Engineer
Conversational AI · NLP · Transformer Experimentation · Full-stack Development
Research interests
Research interest
Intelligent Agents
Evaluation, reasoning, tool use, failure modes, reliability
Related focus areaResearch interest
Retrieval & NLP
Grounded generation, information retrieval, language understanding
Related focus areaResearch interest
Human & Machine Intelligence
Neuroscience applications of ML and neuroscience-inspired intelligent systems
Research interest
Responsible / Explainable AI
Interpretability, transparency, safety, provenance, responsible deployment
Publications
Publication
A Photo-Realistic Simulation and Test Platform for Autonomous Vehicles Research
My first published research project explored machine-learning-based autonomous driving in a photo-realistic simulation environment, connecting steering prediction models with MATLAB and Unreal Engine.
Publication link forthcoming
Detailed Character Bio
LANGUAGES
Python, Rust, TypeScript
FRAMEWORKS
PyTorch, JAX, vLLM, HF Transformers
DOMAINS
LLMs, Interpretability, Alignment, MLOps
I got interested in artificial intelligence by first being interested in human intelligence.
Before I discovered programming and robotics, I planned to study neurology. The career path changed, but the question didn't: what can understanding the way humans think and learn teach us about building intelligent systems?
That curiosity eventually took me from computer engineering into conversational AI and NLP, then into several years of production software engineering at Microsoft, where my work increasingly focused on retrieval, AI infrastructure, agentic systems, evaluation, reliability, and Responsible AI.
Today, I'm pursuing an M.S. in Computer Science at the University of Utah while continuing to build AI systems and explore the research questions behind them.
I tend to gravitate toward messy problems: systems that cross boundaries, models whose failures aren't obvious, and technical questions that require understanding both how something works and why it behaves the way it does.
Outside the technical work, I care a lot about teaching, mentorship, and making technical communities more welcoming — things I've been involved in throughout both university and industry.