Skip to content
YOR / SELECTED ENGINEERING WORK

Projects
with receipts.

Six canonical project worlds. Each case study keeps the work, evidence, contribution boundary, and limitations in the same frame.

06 canonical case studies01 source of truth0 fabricated deployments
01 / THE SELECTED SET

Read the
engineering.

Open a project for the architecture, validation evidence, and limitations. Verified source and live routes remain explicit.

  1. Candidate capability intelligence

    CandidateX

    A provenance-grounded hiring research prototype that turns synthetic candidate documents and selected public technical evidence into role-specific capability estimates, contradiction diagnostics, and targeted interview probes.

    • Python
    • FastAPI
    • Next.js
    • PostgreSQL
    Capability dimensions
    12Canonical technical dimensions documented by the CandidateX scoring framework
    Paper benchmark
    28,800Synthetic candidate-role evaluations reported in the submitted paper across 16 seeds and 6 target roles
  2. Realtime social systems

    Yor Talks V2

    A full-stack social product system built around authenticated conversations, durable persistence, and realtime message delivery.

    • React
    • Vite
    • Express 5
    • Socket.IO
    API checks
    61/61Repository readiness audit; local exercised paths
    Chromium E2E
    17/17Repository readiness audit; not public deployment evidence
  3. Realtime energy intelligence

    Yor Helios

    An experimental FastAPI/WebSocket system that moves energy signals through anomaly analysis into an operator-facing decision surface.

    • Python
    • FastAPI
    • WebSocket
    • Anomaly detection
    System type
    RealtimeFastAPI + WebSocket path for signal updates
    Telemetry boundary
    IllustrativeDeterministic sample data, not field telemetry
  4. Solar planning / decision support

    Yor Zenith

    An experimental solar decision-support experience connecting site and tariff inputs to visual scenarios and financial framing.

    • Next.js
    • TypeScript
    • Three.js
    • Recharts
    Decision model
    Scenario-basedIllustrative values with visible assumptions
    Ownership
    SharedInterface direction separated from architecture/full-stack credit
  5. Texture Forensics / classical image forensics

    Yor AI vs. Real Image Detector

    A classical computer-vision study that uses LBP and GLCM texture features with an SVM to classify AI-generated versus real imagery.

    • Python
    • OpenCV
    • LBP
    • GLCM
    Holdout accuracy
    78.5%107-image holdout; repository-stated evaluation
    Feature representation
    17 valuesLBP + GLCM texture features
  6. Next.js product system

    Personal Developer Portfolio

    A server-rendered engineering portfolio that turns project evidence into a fast, navigable case-study system.

    • Next.js
    • TypeScript
    • React
    • React Three Fiber
    Public project scope
    Registry-backedCanonical CV projects in the recruiter-facing index
    Rendering boundary
    DeferredShared R3F scene loads after idle time; reduced motion retains static artwork