<aside>

ShiftElec is looking for its future experienced researcher to develop non-intrusive electrical consumption monitoring technology in industry and integrate it into industrial optimization systems!

</aside>

Informations pratiques

<aside>

📅 6**-month fixed-term contract**, starting summer 2026

📍 SAMOVAR Laboratory, Télécom SudParis, Institut Polytechnique de Paris.

👤 PhD in AI or Engineer with 3+ years of R&D experience

🖥️ Hybrid: On-site (Belfort-Montbéliard and/or Palaiseau-Evry-Paris and/or Brussels) and remote.

💰 Salary according to profile and experience

🤝 You will work with:

• Faculty Researchers from Télécom SudParis;
• Faculty Researchers from the Energy4Climate center at Polytechnique;
• Faculty Researchers from the Hi!Paris center at Polytechnique & HEC. </aside>

🧑‍💻 The Position

As a Post-doctoral Researcher or Research Engineer, you will be at the heart of technological innovation for the energy transition. You will contribute to developing and testing a first prototype of Non-Intrusive Load Monitoring (NILM) in industrial environments. NILM allows estimating the energy consumption of individual equipment by analyzing only aggregated measurements from the main meter, without installing sensors on each device. Although widely studied in the residential sector, its industrial application remains less developed.

📔 Detailed Program for Your Fixed-Term Contract (8 to 10 months)

You will join the team after preliminary work has been completed by a Data Scientist intern: their mission consists of identifying, extracting, transforming and downloading industrial electrical consumption data; conducting a literature review on NILM; and proposing a first model approach.

Months 1-2: Onboarding, technical specification, development of a first model and validation

• Get familiar with the ShiftElec ecosystem: existing optimization system, database architecture, etc.
• Analyze the structure of the pre-existing optimization system and establish the integration strategy for the NILM model
• Specify the interface contract between the NILM model and the solver: exchanged data formats, complete workflow from power curve retrieval to optimized production schedule calculation
• Support the intern in developing their NILM model:
  • Develop a functional NILM model by exploring different approaches.
  • Train and optimize the model on industrial data.
  • Validate absolute performance.
  • Write a scientific paper on your methodology and results.

Months 3-4: Model improvement and optimization

• Test and compare different models.
• Optimize performance.
• Document attempts.

Months 5-6: System integration

• Develop the integration pipeline between the NILM model and the optimization system.