JD.

Applied AI · Systems Thinking · Production Code

Jane
Doe.

I build AI-powered systems that ship — from agentic workflows and LLM integrations to full-stack web applications. Physics background. Production mindset.

View ProjectsGet In Touch
Scroll
01 / About

Rigorous by training.
Builder by instinct.

I hold a double major in Computer Science and Physics and a Master's degree in Physics — a background that trained me to model complex systems, reason carefully about tradeoffs, and stay skeptical of easy answers.

Today I build AI-powered applications: agentic systems, LLM integrations, and automation pipelines that solve real problems. My current work includes AI model evaluation at Mercor, where I write and assess physics-grounded prompts to improve reasoning capabilities in frontier models.

I'm drawn to roles where AI is a force multiplier — not a buzzword — and where careful engineering decisions compound over time.

AI & ML

OpenAI APIAnthropic APILangChainPrompt EngineeringRLHF / Model EvaluationAgentic Workflows

Development

ReactNext.jsTypeScriptPythonPHPREST APIs

Tools & Infra

VercelAWSGitHubDockerMoodle / LMSTailwind CSS
02 / Projects

Selected work — built, shipped, and documented.

01

Physics Reasoning Evaluator

PythonAnthropic APIRLHF

A structured evaluation harness for testing LLM performance on physics problems — from kinematics to quantum mechanics. Generates adversarial prompts, scores chain-of-thought reasoning, and surfaces failure modes in model responses. Built during model evaluation work at Mercor.

  • →Systematic prompt taxonomy across 6 physics domains
  • →Automated scoring pipeline with rubric-based grading
  • →Identified 3 consistent reasoning failure modes in tested models
02

Document Q&A Agent

Next.jsOpenAI APILangChainVercel

A production-deployed RAG application that lets users upload PDFs and ask questions against them. Built with a streaming response UI, citation tracking, and chunk-level retrieval scoring. Demonstrates full-stack AI development from API integration to deployed UI.

  • →Streaming responses with real-time citation display
  • →Chunk relevance scoring surfaced to the user
  • →Deployed to Vercel with edge functions for low latency
03

Agentic Workflow Prototype

PythonLangChainOpenAIREST APIs

A multi-step AI agent that autonomously gathers, summarizes, and routes information across external APIs. Demonstrates tool use, memory management, and graceful error handling in an agentic loop. Designed to showcase build-vs-buy reasoning with documented architecture decisions.

  • →Tool-use with 4 external API integrations
  • →Structured decision log for every agent action
  • →Handles partial failures with fallback routing
03 / Experience

2023 – Present

Part-time / Contract

AI Model Evaluator

Mercor

Develop and assess physics-based prompts to evaluate reasoning quality in frontier language models. Work includes constructing adversarial test cases, scoring chain-of-thought responses against rubrics, and identifying systematic failure modes — contributing directly to RLHF data pipelines.

Prompt EngineeringModel EvaluationPhysics ReasoningRLHF

2022 – 2024

Full-time

Educational Technologist

Austin College — IT Department

Sole administrator for the college's Moodle LMS — responsible for platform maintenance, faculty training, and integrations with institutional systems. Also maintained 3D printer lab, provided tier-2 technical support, and documented institutional IT processes.

Moodle / LMSSystems AdministrationTechnical SupportDocumentation

2019 – 2022

Academic

M.S. Physics

Graduate Studies

Advanced study in computational and theoretical physics. Developed strong foundations in mathematical modeling, simulation, data analysis, and technical writing. Thesis research required designing experiments, building analysis pipelines, and presenting results to mixed audiences.

Computational PhysicsData AnalysisPythonLaTeXResearch Methodology

04 / Contact

Let's build something worth building.

I'm actively looking for applied AI development roles where careful engineering and genuine curiosity matter. If that sounds like your team, I'd love to talk.

Say Hello
GitHub ↗LinkedIn ↗