Robel Mezgebe

I'm a Computer Science graduate from Cleveland State University who gravitates toward the parts of a system most people never see — the data pipelines, the query plans, the scheduler. I've shipped full-stack features for a marketing intelligence platform, cut API latency by rewriting raw SQL, and spent a semester inside the Pintos kernel implementing priority scheduling and synchronization primitives from scratch. Lately I've been building retrieval-augmented generation systems on PostgreSQL and pgvector.

1. Oct 2025 — Present

   Software Engineer Intern, Full-Stack

   OmniOS · Remote

   Building the data layer behind a marketing intelligence platform, from ingestion pipelines through to AI-powered search.

   • Engineered server-side data tracking pipelines that capture high-fidelity product metrics and make them queryable downstream.
   • Implemented retrieval-augmented generation features on PostgreSQL and pgvector, delivering fast contextual embedding search over customer data.
   • Architected full-stack functionality in Node.js and PostgreSQL, optimizing workflows for complex data ingestion.

   • Node.js
   • PostgreSQL
   • pgvector
   • RAG
   • REST APIs

2. Feb 2026 — May 2026

   Software Engineering Intern

   StoveGuard / 1 Group LLC · Middleburg Heights, OH

   Built internal tooling for enterprise product and inventory data, working directly with the stakeholders who used it.

   • Developed internal data platform tools in React and Node.js to manage workflows across product and inventory datasets.
   • Reduced API latency significantly by profiling and rewriting complex raw SQL queries.
   • Translated business requirements into technical specifications and shipped maintainable features to production.

   • React
   • Node.js
   • SQL
   • PostgreSQL

• NYC Taxi Tip Predictor

  2026

  An end-to-end ML pipeline serving predictions over 100K+ records.

  Built ingestion, preprocessing, and training with pandas and XGBoost, then containerized each component with Docker and served inference through a FastAPI microservice. Decoupling training from inference meant models could be retrained without touching the serving path.

  • Python
  • XGBoost
  • pandas
  • FastAPI
  • Docker

• Pintos Kernel & TinyFS

  2025

  Operating system internals, written from scratch in C.

  Modified the Pintos kernel to implement priority scheduling and synchronization primitives including semaphores and locks. Separately designed TinyFS, a lightweight file system in C that behaves deterministically over a simulated virtual disk abstraction.

  • C
  • Linux
  • Concurrency
  • File Systems

• NASA Psyche Mission Card Game

  2026

  An educational card game about NASA's Psyche mission.

  Built gameplay mechanics and UI state management in object-oriented C# on Unity, then tuned asset rendering and resource management to stay performant in WebGL. Shipped a cross-platform browser build and presented the work at a university symposium.

  • Unity
  • C#
  • WebGL

Languages

• Python
• C
• Java
• JavaScript
• C#
• SQL

Data & AI

• Machine Learning
• RAG Pipelines
• pgvector
• scikit-learn
• XGBoost
• Feature Engineering

Backend

• FastAPI
• Node.js
• Express.js
• Prisma
• REST APIs
• React

Infrastructure

• Docker
• PostgreSQL
• Git & GitHub
• CI/CD
• Cloud

1. May 2026

   B.S. Computer Science

   Cleveland State University · Cleveland, OH

   GPA 3.4 · Data Structures & Algorithms, Operating Systems, Database Systems, Computer Networks, Real-Time Embedded Systems

2. Fall 2023

   Associate of Science

   Cuyahoga Community College · Highland Hills, OH

   GPA 3.54 · Phi Theta Kappa Honor Society