Shpetim Gashi
Data Scientist, Forecasting & Optimisation
- Lienz, Austria
- contact@shpetimgashi.com
- GitHub
Profile
Data science and software development at HELLA InnovationLAB from 2022 to 2026, from dual study to project lead of a reinforcement-learning research project. Python for time-series forecasting, linear optimisation and reinforcement learning, plus the tested software around the models: selected users worked productively with a material optimisation tool I further developed before I left.
Part-time M.Sc. in Data Science and Intelligent Analytics, weekend format, expected July 2027, attendable remotely and compatible with full-time employment. Available at short notice and open to relocation.
Skills
Programming Languages
- Python
- TypeScript
- JavaScript
- SQL
Machine Learning & AI
- Time-Series Forecasting
- Demand Forecasting
- SARIMA
- SARIMAX
- statsmodels
- LSTM
- Reinforcement Learning (PPO)
- TensorFlow
- PyTorch
- scikit-learn
- pandas
- NumPy
- Feature Engineering
- Linear Programming (PuLP)
- Building Energy Modelling (R-C)
Databases & Data
- PostgreSQL
- SQLAlchemy
- SQLite
- Data Pipelines
- Power BI
- Data Visualisation
Backend
- REST integration
- Flask
- Node.js
- Streamlit
Frontend
- React
- Next.js
Testing
- pytest
- pytest-cov
- Playwright
- Vitest
DevOps & Tooling
- Git / GitLab
- CI/CD
- Docker
- Linux
- Jira
- Confluence
- Scrum
- GitHub Issues
- AI-assisted development (coding agents)
Spoken Languages
- German (C2)
- English (C1/C2)
- Albanian (C1)
- French (B2)
Experience
Data Science Projects & Postgraduate Study
Jun 2026 - PresentAvailable at short notice for full-time work. Own projects in deep learning and software testing alongside a weekend M.Sc. in Data Science and Intelligent Analytics (expected July 2027).
- Designed and evaluated a deep-learning project on predictive maintenance (MLP, CNN, transfer learning; TensorFlow/Keras): derived the achievable error bound before modelling and showed the best network matched a four-parameter linear regression, built with AI coding agents.
- Built two web projects end to end, tested in CI and not yet launched (TypeScript, Next.js, Astro, Playwright, CI/CD), one turning photographs into printable 3D meshes and G-code.
- Built a mutation-testing harness that re-injects past defects; its first run showed an authorisation bypass would pass type checks, unit tests and the gate's own smoke test, and I fixed the test.
Project Lead, Data Science & Software Development
Sept 2025 - May 2026Hands-on project lead: led the HELLA-side workstream of a reinforcement-learning research project, built its simulation pipeline in Python, and further developed an optimisation tool that selected users worked with productively.
- Led the HELLA-side workstream of the cooperative research project "Reinforcement Learning for Comfort Conditions in Buildings": delegated work, steered technical direction, accountable for outcomes, progress, budget and reporting; ran delivery in sprints with Jira and Confluence.
- Introduced AI tools to colleagues who were technical but not developers, in weekly team meetings with 1:1 follow-ups; they went on to use them independently for data analysis, research and documentation.
- Further developed a material optimisation tool built on a colleague's linear-programming model (Python/PuLP, inputs from SAP): extended the model and built its Streamlit and Flask interface: results with KPIs and an overview by material group, plus a KPI history for evaluation.
- Added pytest tests and a CI pipeline that checks reference cases with known expected results before every release build; selected users worked with the tool productively before I left.
- Built the simulation-based experimentation pipeline in Python: PPO agents (stable-baselines3) in a custom Gymnasium environment backed by Radiance daylight simulation, computed on demand and cached, tracked in TensorBoard, with evaluation criteria defined before any results were in.
- Brought a university software project to production readiness as sole developer: Flask with its relational data model in PostgreSQL (SQLAlchemy), authentication and REST integration, tested with pytest and Playwright, shipped with Docker and GitLab CI/CD.
- Contributed to a customer-facing product configuration and AR platform (Next.js, React, TypeScript, Node.js): broke features down into technical tasks, was responsible for the UI/UX workflows, and worked on database-backed features and CI/CD.
Data Science & Software Development (Dual Study Programme)
Sept 2022 - Aug 2025Dual study programme combining an engineering degree at MCI with applied R&D work across InnovationLAB, Product Management and international sales.
- Built a time-series forecasting pipeline benchmarking (S)ARIMA baselines against LSTM models (TensorFlow/Keras) to test whether the added complexity paid off.
- Explained to engineering, Product Management and Sales what the forecasting models could support and where their limits were.
- Built Power BI dashboards that turned operational data into reporting non-technical stakeholders could act on.
- Built Python web-data collection workflows with Selenium and requests for structured downstream analysis.
- Managed a software project with Product Management, from requirements engineering to coordinated implementation: a web application providing product documents, in effect an early digital-product-passport MVP.
- Bachelor thesis: *Optimization of Solar Shading Systems: A Comparison of Machine Learning and R-C Models for Reducing Energy Demand in Buildings*; built the R-C and LSTM/RNN models in Python/TensorFlow and designed the benchmarking setup that compared them on equal terms.
Student Representative
Sept 2022 - Jun 2026Elected representative at programme and institutional level across two universities, alongside study and work.
- Chair of the Student Representation at MCI until 07/2025: set its agendas and documented its decisions.
- Represented student interests to faculty and institutional leadership.
Working Student, SAP Data Migration
Jan 2022 - Oct 2022Part-time working-student role.
- Manually transferred large volumes of data into SAP, from legacy-system entries to scans and paper documents.
- Onboarded new working students and checked the quality of their entries.
Technical Consultant
Oct 2020 - Sept 2021Customer-facing technical consultancy, from requirements on site to documentation for non-technical users.
- Set up and troubleshot home IT and networks at customers' homes, analysing requirements on site.
- Produced documentation for non-technical users; 9.6/10 average customer satisfaction.
Salesperson / Team Assistant
Mar 2013 - Oct 2020Seven years in retail as salesperson and team assistant.
- Supported KPI monitoring and operational planning.
- Operationally responsible for a restructuring project that reduced costs by 10% and increased seasonal sales by 25%.
Selected Projects
Machine learning & data science
Simulation-based reinforcement-learning experiments for control problems in building technology. PPO agents trained with stable-baselines3 and Gymnasium, Radiance-based simulation tooling for the environment, and TensorBoard for experiment tracking, with simulation assumptions and limitations kept explicit.
Material optimisation tool built on a colleague's linear-programming model in PuLP, which I further developed: extended the model, built its Streamlit and Flask interface and pytest tests, and built a CI pipeline that checks reference cases with known expected results before every release build. Selected users worked with it productively before I left.
Time-series forecasting pipeline with the evaluation fixed before the models: LSTM models in TensorFlow/Keras benchmarked against (S)ARIMA baselines in statsmodels, with rolling evaluation windows and each horizon scored against the baseline.
Graded M.Sc. deep learning project predicting an engine's remaining useful life from temperature readings. Grouped cross-validation to prevent data leakage, a measured noise floor, and a clear result: a four-parameter linear model matched the neural network.
Software engineering
A university software-engineering project built around a real-world building-technology use case: a maintainable Flask application for collecting and managing structured user feedback, covering authentication and role handling, PostgreSQL persistence, administrative workflows, REST integration, automated testing, containerisation and CI/CD.
Customer-facing platform for configuring products and visualising them in augmented reality. Broke features down into technical tasks, was responsible for the UI/UX workflows, and contributed database-backed features and CI/CD in a TypeScript/Node stack.
This site. A single-source-of-truth architecture where all CV data lives in one schema-validated YAML file that drives the rendered CV, its print version and the project list. Token-based design system with dark and light themes, and a CI pipeline whose checks have to pass before anything is published.
Education
M.Sc. Data Science and Intelligent Analytics
Sept 2025 - expected Jul 2027Part-time weekend programme, attendable remotely and compatible with full-time employment. Expected graduation July 2027.
B.Sc. Smart Building Technologies
Sept 2022 - Jun 2025Certificate, Agile Project Management (Scrum)
Dec 2024 - Jan 2025Certificate, Data Science Summer School
Jul 2023 - Aug 2023