Giorgi Malania

Giorgi Malania

Mechanical & aerospace engineering student at Case Western, and Content Product Manager at General Learning (prev. RevisionDojo, YC F24), where I build the physics curriculum for nearly a million students.

I build things that are physical and things that aren't: CAD and machined parts, curriculum, and software. This fall I'm studying at UC3M in Madrid.

You can reach me at malaniagrgi@gmail.com

Cleveland → Madrid

Projects

oushi

Live

A personal AI assistant built on Gmail and Google Calendar. It processes incoming mail in real time through Google's Pub/Sub push notifications, parses sender and thread context, and surfaces only the messages that require a response.

I designed and deployed the full stack: authentication, the real-time email pipeline, and the daily-brief interface.

  • Next.js
  • TypeScript
  • Supabase
  • Google APIs
  • Pub/Sub
  • Vercel

Yale Hackathon

2nd place

A quantum optimizer for insurance portfolio selection, developed at Yale's YQuantum for the Capgemini and The Hartford challenge. We reduced 50 assets to 8 sectors mapped onto 8 qubits and applied QAOA to identify the optimal allocation under real regulatory constraints.

Circuit-depth optimizations roughly halved the depth and raised simulated accuracy from 17% to 53%, and we validated the result against a brute-force classical solve. Placed 2nd of 35 teams and won Best Computational Accuracy.

  • Python
  • Bloqade
  • QAOA
  • QUBO
  • PyQrack

Engineering

Hurdle Helper

Fall 2025

A training attachment that clamps onto a track hurdle and swings a padded arm into position for trail-leg drills. Designed in SolidWorks, with the loaded arm validated in FEA and refined across several prototype iterations. The mechanism uses a spring-loaded, ball-bearing pivot that returns the arm after each strike.

  • SolidWorks
  • SolidWorks Simulation (FEA)
  • GD&T
  • Engineering Drawings
  • Prototyping

Wind Tunnel (NACA 2415)

Spring 2026

Characterized a NACA 2415 airfoil in a subsonic wind tunnel across seven angles of attack. Surface-pressure data was acquired through LabVIEW and reduced in Python to compute lift and moment coefficients, then benchmarked against vortex-panel-method predictions and published NACA data. Diagnosed and corrected two faulty pressure taps to preserve data integrity.

  • Aerodynamics
  • LabVIEW
  • Python
  • Data Analysis

Optical inspection stage

In progress

A 2-axis gantry that traverses a camera across a part, detects and localizes defects with a YOLO model, and generates structured inspection reports. Modeled on semiconductor inspection workflows; both the mechanical design and the vision pipeline are my own. Currently in progress.

  • Python
  • YOLO
  • OpenCV
  • Creo
  • Ansys