Hello, I'm

Shpetim Gashi

Software Developer & Data Scientist

I deliver end-to-end solutions combining web development and machine learning: database-backed applications (Python backends, React/Next.js frontends) and ML systems (forecasting, reinforcement learning, optimisation). I led a EUR 200,000 AI research project from requirements and reward-function design through validation and rollout. Currently completing an M.Sc. in Data Science and Intelligent Analytics.

View CVView Projects
Professional headshot of Shpetim Gashi

  1. Demand Forecasting Pipeline

    LSTM- and (S)ARIMA-based demand forecasting pipeline for industrial R&D. Rigorous evaluation showed a three-month window at approximately ±5% deviation, informing operational feasibility decisions.

    • Python
    • LSTM
    • TensorFlow
    • SARIMA

  2. RL for Comfort-Based Shading Control

    Proof-of-concept reinforcement-learning pipeline (PPO) for solar-shading control. Reward functions built on Radiance illuminance and daylight-glare criteria; validated feasibility on synthetic simulation data.

    • Python
    • Reinforcement Learning
    • PPO
    • Radiance

  3. Rod-Length Optimisation & Scenario Planning

    PuLP-based optimisation tool translating manufacturing constraints into deterministic cutting recommendations. Extended with scenario planning and a Streamlit/Flask interface; contributed to an estimated 15% material reduction.

    • Python
    • PuLP
    • Streamlit
    • Flask

  4. Vision Planner - Product Configuration Platform

    Full-stack product configuration and AR visualisation platform (vision.hella.info): Next.js/React/TypeScript frontend, Python services, CI/CD pipelines, and security audits. All AI-assisted code reviewed before integration.

    • Next.js
    • React
    • TypeScript
    • Python

All projects →

Get in touch