Skip to content
View alaaeddinethr's full-sized avatar
1 follower · 1 following

Block or report alaaeddinethr

Block user

Prevent this user from interacting with your repositories and sending you notifications. Learn more about blocking users.

You must be logged in to block users.

Content in all repositories owned by your account will be closed.
Maximum 250 characters. Please don’t include any personal information such as legal names or email addresses. Markdown is supported. This note will only be visible to you.
Report abuse

Contact GitHub support about this user’s behavior. Learn more about reporting abuse.

Report abuse
alaaeddinethr/README.md

Hi, I'm Alaa Eddine Tahir

Engineering student at Arts et Métiers (ENSAM), graduating in 2027, with an MSc in Financial Engineering in progress at WorldQuant University. I work where industrial engineering meets data: predictive maintenance, production performance, project and market risk.

Portfolio: alaaeddinethr.github.io · LinkedIn: alaaeddine-tahir · alaa_eddine.tahir@ensam.eu

Data and AI for industry

Project Result Stack
turbofan-rul-prognostics Remaining useful life of aircraft engines (NASA C-MAPSS): 12.3-cycle RMSE with calibrated 80 % intervals XGBoost, PyTorch, Streamlit
Bearing fault detection Vibration diagnosis from bearing physics: 99.8 % on an unseen motor load (CWRU). From a 3rd-place MTI 2025 hackathon project Signal processing, PINN
oee-dashboard OEE, TPM Six Big Losses and Pareto of losses on a production line pandas, Streamlit
Optima Automated CE-marking compliance audit with satisfied / not satisfied / ambiguous verdicts Python, rule engine

Risk and quantitative finance

Project Result Stack
quant-risk-engine VaR backtest on 2019-2024 data: only filtered historical simulation passes the Kupiec test and the Basel green zone NumPy, SciPy
Risk-Engine Construction schedule risk: a most-likely plan has an 8-16 % chance of being met (PERT, CPM, Monte Carlo) NetworkX, SciPy

Every project has tests, continuous integration and results you can reproduce with one command.

Client work

AmbaZ homepage redesign (GESCEM) · SIEFER TRIGONAL® Maroc website · Prospecting CRM

Experience

  • GESCEM (2026): AI layer on a maintenance management system (CMMS) in production; annual maintenance planning engine built on 754 assets and 6,428 real interventions, workload irregularity reduced by 65 %.
  • KA2C Construction (2025): coordination of three trade teams on a building site.
  • Stellantis Metz (2024): performance analysis on an automotive assembly line.
  • Director of AMJE Rabat, Arts et Métiers junior enterprise: team of 8, six client projects.

Stack: Python, SQL, scikit-learn, XGBoost, PyTorch, FastAPI, Docker, Git · CATIA V5, 3DEXPERIENCE, SolidWorks

Pinned Loading

  1. oee-dashboard oee-dashboard Public

    Overall Equipment Effectiveness analytics: OEE per machine and shift, TPM Six Big Losses, MTBF/MTTR and a Pareto of losses to decide where to act first.

    Python

  2. Predictive-Maintenance-Dashboard-Bearing-Fault-Detection Predictive-Maintenance-Dashboard-Bearing-Fault-Detection Public

    Bearing fault diagnosis from vibration signals: 52 physics-based features (FFT, Hilbert envelope, BPFI/BPFO), XGBoost and a physics-informed neural network. 99.8 % on an unseen motor load (CWRU).

    Python

  3. quant-risk-engine quant-risk-engine Public

    Market risk toolkit: VaR and Expected Shortfall (4 methods), out-of-sample backtesting with the Kupiec test and Basel traffic light, Monte Carlo, stress tests, Markowitz.

    Python

  4. turbofan-rul-prognostics turbofan-rul-prognostics Public

    Remaining useful life of aircraft engines (NASA C-MAPSS): 12.3-cycle RMSE with calibrated 80 % prediction intervals. XGBoost, PyTorch, Streamlit.

    Python

  5. Optima Optima Public

    Automated CE-marking compliance audit: extracts regulatory requirements, checks them against a machine product sheet and returns satisfied / not satisfied / ambiguous with the missing evidence.

    Python

  6. Risk-Engine Risk-Engine Public

    Construction schedule and cost risk: Beta-PERT estimates, critical path method and 10,000 Monte Carlo runs. A most-likely plan has an 8-16 % chance of being met.

    Python