Michael Marin
Michael E. Marin, applied AI and machine learning engineer

Building applied AI systems people can test and use

Seeking AI/ML and applied-AI internships and co-ops

I build measurable applied-AI systems across retrieval, evaluation, computer vision, and AI products. My work includes a reproducibly benchmarked RAG platform, a 2025 AI/ML product internship, and leading a five-person NASA PACE capstone team while pursuing an M.S. in Artificial Intelligence at UNT (Expected December 2027).

dallas–fort worth, tx

0

Documents benchmarked

SciFact retrieval evaluation

0+

Test assertions

SIMLYFE engine + LLM modules

0

Person team led

NASA PACE capstone exhibit

0

M.S. AI

expected Dec — UNT, ML concentration

michael@dallas — zsh

$ whoami. ml-engineer — dallas, tx. $ current_focus. → applied AI + ML engineering. → M.S. AI @ UNT (Dec 2027)

$ cat about.md

About Me

I'm Michael, an applied AI builder in Dallas–Fort Worth. I earned a B.S. in Computer Science at the University of North Texas and am now pursuing an M.S. in Artificial Intelligence with a Machine Learning concentration, expected December 2027. I care about the point where an ML idea leaves the notebook and becomes a system people can test, evaluate, and use.

Applied AI and machine learning engineer who builds measurable systems across retrieval, evaluation, computer vision, and AI products. Built RAGOps with reproducible SciFact benchmarking and local-provider load testing; brings complementary 2025 AI/ML product internship experience, NASA PACE computer vision leadership, and a shipped AI-assisted life simulator.

Selected work

Flagship Projects

A measured retrieval platform, a public computer-vision exhibit, and a shipped AI-assisted product.

01 / Personal Applied-AI / Retrieval Engineering Project · 2026

RAGOps — Production-Style RAG Platform

The problem

A RAG system is difficult to improve when ingestion, retrieval quality, latency, citations, cache behavior, and failures are hidden behind one answer endpoint. The goal was to build an operations-oriented platform where…

The build

Built a production-style RAG platform with FastAPI, PostgreSQL/pgvector, Redis, asynchronous ingestion, hybrid retrieval, reranking, Prometheus metrics, and a React operations console.

Python
FastAPI
PostgreSQL
pgvector
Redis
BM25
+8 more
Read the case study
RRF led the held-out benchmark

0.8059

Recall@10

0.6816

nDCG@10

  • → 5,183 SciFact documents · 300 held-out queries
  • → 24.56 ms p50 · 31.75 ms p95 · 9,234 local-provider requests
  • → Cross-encoder retained as a slower, lower-quality negative result
Repository

02 / NASA PACE · University of Maryland IMD · University of North Texas · Jan 2024 – May 2025

WAVE: From Space to Ocean (NASA PACE Capstone)

The problem

NASA's PACE mission produces complex oceanographic and atmospheric satellite data. The capstone challenge was to make that science approachable through a real-time, gesture-driven public exhibit.

The build

Led a 5-person team building a real-time computer vision and interactive visualization system for NASA PACE satellite data.

C#
C++
Python
Unity
OpenCV
ZED SDK
+6 more
Read the case study
Kennedy Center, Washington, D.C.
  • → Premiered at the Kennedy Center in March 2025
  • → Led a 5-person capstone team
  • → Computer vision, stereo cameras, and interactive rendering

03 / Personal Project · 2025 – 2026

SIMLYFE: AI-Assisted Life Simulator

The problem

LLM-generated content is compelling but unreliable: left unconstrained it breaks game-state coherence, and calling it naively from the browser leaks API keys. The challenge was building a playable, shippable game where…

The build

Built and shipped a mobile-first, browser-based life simulation game where players age a character year by year through careers, relationships, finances, and procedurally generated life events.

React 19
JavaScript
Vite 8
Supabase Edge Functions
Firebase Auth
Firestore
+3 more
Read the case study
Live in production
Live DemoRepository

350+ automated test assertions · Vitest

Interactive ML Lab

Real algorithms running in your browser — not videos.

Real in-browser inference — a convolutional net served with onnxruntime-web, the same pattern used to ship ML to the edge.

Track record

Experience & Projects

  1. 2025 – 2026

    SIMLYFE: AI-Assisted Life Simulator

    Personal Project

    • Built and shipped a mobile-first, browser-based life simulation game where players age a character year by year through careers, relationships, finances, and procedurally generated life events.
    • Separated deterministic game systems from authenticated OpenAI narrative generation behind a Supabase Edge Function.
    • Validated bounded requests and model output at the server boundary while keeping provider credentials out of the browser.
    Details

  2. 2026

    RAGOps — Production-Style RAG Platform

    Personal Applied-AI / Retrieval Engineering Project

    • Built a production-style RAG platform with FastAPI, PostgreSQL/pgvector, Redis, asynchronous ingestion, hybrid retrieval, reranking, Prometheus metrics, and a React operations console.
    • Benchmarked BM25, all-MiniLM-L6-v2 dense retrieval, RRF, and a cross-encoder trained only on SciFact's disjoint training split across 5,183 documents and 300 held-out queries; RRF led with 0.8059 Recall@10 and 0.6816 nDCG@10.
    • Built reproducible evaluation and load-testing pipelines with versioned datasets, artifact checksums, per-query traces, and CI quality gates; recorded 24.56 ms p50 and 31.75 ms p95 over 9,234 successful requests to the local-provider stack.
    Details

  3. 2025

    Product Analyst Intern, AI/ML team

    Mr. Cooper Group

    • Supported agentic AI products for customer service automation across Azure DevOps and MySQL.
    • Participated in product experimentation, user validation, and AI feature evaluation activities.
    • Collaborated with product, engineering, and business stakeholders on AI and machine learning initiatives.
    Details

  4. Jan 2024 – May 2025

    WAVE: From Space to Ocean (NASA PACE Capstone)

    NASA PACE · University of Maryland IMD · University of North Texas

    • Led a 5-person team building a real-time computer vision and interactive visualization system for NASA PACE satellite data.
    • Developed gesture-recognition and multi-camera tracking workflows using OpenCV, CUDA, Unity, and ZED stereo cameras.
    • Built interactive data visualizations for real-time oceanographic satellite analytics.
    Details

Capabilities

Technical Skills

Evidence behind the stack

Capabilities tied to shipped work

Retrieval engineering and evaluation

RAGOps benchmarked 5,183 SciFact documents and 300 held-out queries; RRF led with 0.8059 Recall@10 and 0.6816 nDCG@10.

AI product evaluation

2025 Product Analyst Intern on the Mr. Cooper AI/ML team: user validation, feature evaluation, and cross-functional delivery.

Computer vision

Lead Developer for a 5-person NASA PACE capstone team using OpenCV, CUDA, Unity, and ZED stereo cameras.

LLM systems

SIMLYFE separates deterministic simulation logic from authenticated, schema-validated OpenAI event generation.

Cloud engineering

AWS Machine Learning — Specialty and Google Cloud Professional Cloud Architect certifications.

Programming Languages

Python
TypeScript / JavaScript
C++
SQL

AI/ML & Data Science

PyTorch
TensorFlow
OpenCV
NLP
RAG Pipelines
BM25
RRF
Dense Retrieval
Cross-encoder Reranking
Retrieval Evaluation
LLM Evaluation
Scikit-learn

Cloud & DevOps

AWS
Google Cloud (Vertex AI, BigQuery)
Docker
Kubernetes
GitHub Actions
CI/CD pipelines
Prometheus
Grafana
Linux

Backend

FastAPI
Node.js
PostgreSQL
pgvector
Redis
Asynchronous Ingestion
Flask
Firebase
RESTful APIs

Frontend

React
Next.js
Tailwind CSS
Data Visualization

Academic foundation

Education

Pursuing a master's in AI while shipping production systems.

M.S. AIML concentration
Expected December 2027

M.S., Artificial Intelligence (Machine Learning concentration)

The University of North Texas (UNT)

Selected graduate coursework

Applied Artificial Intelligence
Advanced Machine Learning
Graduated · May 2025

B.S., Computer Science

The University of North Texas (UNT)

GPA 3.25 (4.0 Scale)

Cloud credentials

  • AWS Certified Machine Learning — Specialty

    Amazon Web Services · Issued June 2025

  • Google Cloud Certified Professional Cloud Architect

    Google Cloud · Issued October 2025

Third-party proof of the NASA PACE exhibit

Independent Coverage

UMD College of Computer, Mathematical & Natural Sciences

WAVE exhibit premieres at the Kennedy Center

Coverage of the NASA PACE 'Wave: From Space to Ocean' exhibit premiere in Washington, D.C.

Read coverage →

UMD Immersive Media Design

WAVE: New understanding about Earth's oceans and atmosphere

Feature on the interactive satellite-data exhibit built with the NASA PACE team and UNT capstone students.

Read coverage →

NASA PACE Mission

WAVE: From Space to Ocean

The official NASA PACE mission page for the WAVE exhibit.

Read coverage →

auto-updates every 6h

Live from GitHub

OPENSOURCE_ALLDATA

—

updated Oct 2026

xmike04.github.io

Evidence-first applied AI & ML engineering portfolio: computer vision, LLM systems, AI product evaluation, and cloud engineering.

TypeScriptupdated Oct 2026

xmike04

Applied AI & ML engineering profile for Michael Marin.

updated Oct 2026

RAG_OPS

Production-style RAG platform with FastAPI, pgvector, Redis, async ingestion, hybrid retrieval, reproducible SciFact evaluation, Prometheus, and a React ops console.

Pythonupdated Oct 2026

Pacheco-Taco

—

CSSupdated Oct 2026

SIMLYFE

AI-assisted life simulator with deterministic game logic, authenticated OpenAI events, Firebase cloud saves, and 350+ Vitest assertions.

JavaScriptupdated Oct 2026

Get in touch

Let's build production AI together

Seeking AI/ML and applied-AI internships and co-ops. Based in Dallas–Fort Worth and open to strong applied-AI teams.

Dallas–Fort Worth, TX