STAGE - Ingénieur Algorithmie et logiciel - F/H

What if your internship made a real impact?

Ranked #1 most sustainable company in the world and 2nd in the HappyTrainees ranking, Schneider Electric offers a rewarding experience, recommended by 92.9% of our interns and apprentices.

And that's no coincidence! Our culture is built on strong values: Inclusion, Mastery, Purpose, Action, Curiosity, and Teamwork, enabling everyone to fully contribute to the transformation toward a more sustainable world.

You'll be joining a company with 150,000 employees, present in over 100 countries, and a global leader in energy management and automation.

Joining Schneider Electric means much more than just becoming part of a large international group!

Context:

IoTsensor is a human-sized team (15 people) based at the Intencity site in Grenoble.

We design smart sensors.

Our expertise includes:


* Sensitive elements and their characterization


* Electronic and algorithmic design


* Data acquisition campaigns


* Scientific analysis and technological innovation

As a reference team within Schneider Electric, we support other entities in their technological choices.

Our projects span various environments (industrial, residential, building), with strong expertise in people counting and ultra-low power.

Internship Objective

As part of our work on mmWave radars, we are developing products and exploring their potential to:


* Integrate this technology into Schneider products


* Innovate in detection algorithms and filtering out false signals

The goal of the internship is to design algorithms capable of extracting quality-of-life indicators from radar data: activity level, sleep quality, lifestyle habits, etc., and to improve detection quality by reducing false positives.

Your Missions:

The order of tasks is indicative and may evolve depending on project progress.

mmWave Radar Data Acquisition:


* Use existing internal data and algorithmic needs to define data requirements (scene types, data types, granularity, recording architecture)


* Conduct and supervise a data recording campaign (max 2 weeks)


* Visualize and analyze the data to understand its nature and limitations

Defining Algorithm Improvement Areas:

(ML, neural networks, or classical algorithms depending on skills and selected topics)


* Conduct a literature review on human pattern detection algorithms (occupancy, activity, sleep, routines)


* Study methods for rejecting false detections (moving objects, noise, position)


* Define algorithmic approaches (ML, neural networks, classical methods) based on skills and needs


* Design appropriate cost functions and evaluation methods

Proof of Concept Implementation:

(Python, C++, or embedded depending on profile)


* Implement a POC for human pattern detection


* Implement a POC for false detection rejection


* Experiment, test, and analyze performance

Validation and Conclusion:


* Evaluate algorithm performance


* Write a summary...