Ohinoyi Moiza

CS & Math student at Texas Tech University

Co-founder & CEO at SELT. Building production-ready applications with modern JavaScript frameworks.

GitHub LinkedIn Email

Experience

National Technology Consulting Intern, AI & Data

Jun 2026 – Aug 2026

Ernst & Young LLP (EY Americas) · Houston, TX

Worked on an enterprise knowledge graph pipeline that turned decades of unstructured capital project documents into structured, searchable data for a Fortune 10 energy client.
Built testing and validation workflows and evaluated data and retrieval quality across Neo4j, Azure AI Search, Azure Data Factory, and Snowflake.
Found a classification issue causing graph relationships to receive generic labels, traced it through the pipeline, and helped test the fix.

National Technology Consulting Intern, ServiceNow/BPM

Jun 2025 – Aug 2025

Ernst & Young LLP (EY Americas) · Houston, TX

Built an agentic AI proof of concept for insurance underwriting that automated document intake, risk scoring, case creation, and policy generation.
Evaluated GPT-4o-mini, Llama 3, Claude, and Gemini across underwriting tasks like risk scoring and fraud detection, and recommended the best-fit model.
Developed the underwriting agent in Python, working on its architecture, data contracts, validation controls, and end-to-end demo.

Research

Patient-Specific Seizure Propagation Modeling for Epilepsy Surgery Planning

Honors College URS Program

Dr. Emily Pereira's Lab, Dept. of Electrical & Computer Engineering, Texas Tech University

Built fractional-order simulation code using Grünwald-Letnikov derivatives to model seizure propagation from real intracranial EEG data, processing 262 electrodes across 3 patients from the UPenn OpenNeuro dataset. Ran simulations across multiple parameter values, built interactive 3D visualizations, and deployed the workflow on TTU's HPC cluster via Globus Compute.

Presented at the Texas Tech Undergraduate Research Conference, April 2026
Top 10 finalist, URC Commercialization Competition, April 2026 (CortexCompass)
Lightning talk, GlobusWorld 2026 (NSF-funded travel)
Poster presentation, CRA UR2PhD Workshop, New Orleans, April 2026

View project on GitHub →

Education

Texas Tech University

Expected May 2027

Bachelor of Science, Computer Science & Mathematics · Lubbock, TX

GPA: 3.7/4.0 · Honors College · 6× Dean's List

Relevant coursework: Data Structures & Algorithms, Operating Systems, Object-Oriented Programming, Software Engineering, Computer Architecture

Awards & Certifications

JPMorgan Code for Good · 1st Place
Google × MLH AI Roadshow · Top 3 Nationally
Texas Tech HackWesTX · 2nd Place
YC Startup School S26 · Top 5% (1 of 6,000 selected from 30,000+ applicants)
Google Software Engineering Program
Salesforce Futureforce Summit · 1 of 100 selected

Skills

Languages

JavaScript, TypeScript, Java, Python, Go, C++, SQL

Frontend

React, Next.js, React Native, Vue.js, Redux, Tailwind CSS

Backend

Node.js, Express.js, REST APIs, gRPC, Microservices

AI & Data

LLMs, RAG, GraphRAG, Knowledge Graphs, Pandas, NumPy, Neo4j, Snowflake

Cloud & Infrastructure

Azure, Azure AI Search, Azure Data Factory, Docker, Linux, CI/CD, Globus Compute

Databases

MongoDB, MySQL, Redis, Oracle

Tools

Git, GitHub, Tableau, Power BI

Projects

Pulse

Device telemetry ingest service with a bounded queue for 429 load shedding, per-device ring buffers, and non-blocking WebSocket fan-out.

View on GitHub →

Flux

Python

Real-time image generation gateway with backpressure handling and latency tracking over WebSocket.

View on GitHub →

Quant Volatility Research

Python

Neural network calibration of stochastic volatility models using real SPX options data.

View on GitHub →

Career Command

TypeScript

AI-powered career management platform that helps students organize job applications, track opportunities, and manage their recruiting process.

View on GitHub →

Anchor

Python

Confidence-aware reconstruction system for wearable sensor data. Fills sensor dropouts in wrist-worn fitness tracking data, scores confidence in each reconstruction, and uses that score to gate whether a rep counts toward a personal record.

View on GitHub →