AI systems architect. 8.5 years: statistics to LLMs. Sole architect since 2022.
I love systems — the logic underneath them, the order they create out of chaos.
Across 8.5 years — statistics, deep learning, real-time decision systems, LLM and agentic orchestration — the work underneath has stayed the same: design the structure, own the outcome. Sole architect on every system I've shipped since 2022. I'm past wanting someone else's small piece of a system. I want the whole thing — the architecture, the team, the result.
Who. Yeriko Vargas — AI systems architect, 8.5 years in data: statistics, machine learning, deep learning, LLM and agentic systems.
What. Sole architect and builder on every system shipped since 2022 — RAG, agent frameworks, real-time decision systems, forecasting.
Where. Detroit, Michigan. Remote-first, has worked contract, hybrid and remote roles across automotive, finance, hospitality and distribution.
Reach. ygvargas93@gmail.com · (646) 771-6111 · linkedin.com/in/yeriko-vargas
Sole architect and builder of a company's AI infrastructure from zero: a training hub and multi-mode assistant on LLMs, RAG, and agent workflows that replaced four spreadsheets and a paper sign-up process, company-wide. I own it end to end — registration, training, attendance, retention, knowledge retrieval, natural-language access to business data. Still running, still mine.
A production retrieval and agent framework built for financial decision support, with telemetry, evaluation, and monitoring built into every step — not just prompts stitched together, architecture. Paired with LSTM forecasting systems feeding the same decisions.
A system that listens to a live venue and adapts in real time: 100+ DSP features through PCA and KMeans into similarity scoring, backed by a continuous evaluation loop. Running in production.
| Role & employer | Dates · type | What I did | |
|---|---|---|---|
| Artificial Intelligence Engineer M&A Supply Company | Mar 2026 – Aug 2026 · Contract, Remote | Sole architect of the company's AI infrastructure — see the Dealer Training Hub project above. | 6 mo |
| Founder & Principal Systems Architect Custom AI Models | Jun 2022 – Feb 2026 · Self-employed | Built and ran a one-person AI consultancy: production RAG and agent systems for financial decision support, plus a real-time audio intelligence platform — see Projects above. | 3 yr 8 mo |
| Lead Data Scientist Ford Motor Company | Mar 2022 – May 2022 · Contract, Remote | Analyzed large-scale automotive datasets (CVDOS) to support engineering decisions. Built automated pipelines across SQL and Hadoop, cutting manual reporting errors. Applied unsupervised ML to surface recurring vehicle defect patterns, on AWS SageMaker. | 3 mo |
| Data Scientist Soothsayer Analytics | Mar 2021 – Aug 2021 · Full-time | Built fraud-detection and customer-segmentation models across automotive and insurance data. Applied Random Forest, clustering, and anomaly detection to surface high-risk service locations, with SHAP to make the findings business-readable. | 6 mo |
| Data Analyst Official Driving School | Sep 2020 – Feb 2021 · Contract | Built the organization's first AI-ready data infrastructure from raw operational data, with R-Keras forecasting models for private-lesson demand across 50+ locations. | 6 mo |
| Deep Learning Analyst, Automotive Profit Optimization FCA Fiat Chrysler Automobiles (Stellantis) | Oct 2018 – Aug 2020 · Contract, Hybrid | Built deep learning models across ~250K vehicle configurations to optimize profitability, reducing 108 predictors to 14 key drivers — insights that shaped real configuration decisions. | 1 yr 10 mo |
| Biostatistician Beaumont Health | Jul 2018 – Sep 2018 · Internship | Led biostatistical analysis on longitudinal cardiovascular data — mixed-effects models and survival analysis to quantify patient-level risk. | 3 mo |
| Applied Machine Learning Researcher Oakland University Mathematics and Statistics | Jan 2018 – Jun 2018 · Part-time | Applied ML and statistical research across 100M+ row datasets spanning economics and engineering, with scalable pipelines in Python and R. | 6 mo |
One mark. One thread, eight years long — from a statistics degree to owning the whole system.
From a statistics foundation to owning full systems end to end.
Statistics at Oakland, biostatistics at Beaumont Health, applied ML research — the rigor everything since is built on.
Deep learning at FCA, fraud and segmentation models at Soothsayer, automotive data science at Ford — shipping ML and DL that held up at scale.
Sole architect since 2022: RAG, agent frameworks, real-time audio intelligence, the Dealer Training Hub — owning the whole system, not a piece of it.
Eight and a half years. Every stage shipped, nothing left half-built.
A statistics foundation — a master's and bachelor's in Applied Statistics at Oakland University — underneath eight and a half years of shipped systems in deep learning, forecasting, agentic architecture and real-time intelligence.
MS, Applied Statistics — Oakland University, 2019
BS, Applied Statistics — Oakland University, 2017
Automotive, finance, distribution, hospitality, healthcare.
Detroit, MI · remote-first. I read every message myself.
Email me(646) 771-6111 · linkedin.com/in/yeriko-vargas
Yeriko Vargas