Image Processing Intern: AI-Assisted Tools for CalVal

Loft Orbital; Toulouse; On-site; EUR 13,000 - 16,000; Full time

Benefits offered by this job

Gratification: €1,300 gross/month

Lunch vouchers

Sustainable mobility package

International environment

Mentoring program

Job summary

Loft Orbital is seeking an intern to join the End-to-End Validation & Performance (EVP) team in Toulouse. You will contribute to image-processing, calibration/validation activities for Earth-observation missions and turn work into an evaluated engineering prototype.

The internship includes a 6-month duration with an early-2027 start, based in the city center. You will gain experience in radiometric calibration, PRNU/DSNU correction, experimental design, and building reproducible engineering

Qualifications

• Student in engineering, applied physics, remote sensing, image processing, data science, or a related field.
• Fundamentals of signal or image processing and a strong interest in Earth observation.
• Basic programming skills in Python and comfort working with scientific data.
• Analytical mindset, rigor, and ability to document assumptions and limitations.
• Interest in calibration, validation, sensor behavior, or image quality assessment.
• Exposure to satellite imagery or remote-sensing data.
• Experience with optimization, statistical estimation, or machine-learning evaluation.
• Familiarity with image-analysis libraries and scientific Python tools.
• Understanding of radiometric calibration or data-quality assessment.

Responsibilities

• Review DSNU, radiometric calibration, and EVP processing/CAL/VAL practices.
• Collaborate with EVP engineers to identify suitable images, metadata, and reference data.
• Document data limitations and provenance.
• Prepare a reproducible dataset and establish a holdout evaluation strategy.
• Implement a baseline estimator: a data-fit term with a regularizing prior.
• Explore AI/ML methods only where data are available and compare to the baseline.
• Evaluate correction quality, residual fixed-pattern structure, stability across images, runtime, and failure cases.
• Package the workflow as a prototype script or notebook with plots, quality checks, and an auditable record of inputs, parameters, and outputs.
• Present findings and recommendations to EVP and CAL/VAL stakeholders; document evidence needed before operational use.

Skills

Python basics

Image processing

Remote sensing

Data science

Analytical mindset

Documentation & reporting

Optimization/ML evaluation

Image-analysis libraries

Education

Student in engineering / applied physics / remote sensing / image processing / data science

Job description

Wanna join the adventure?

Loft Orbital is looking for an intern to join the End-to-End Validation & Performance (EVP) team and contribute to image-processing and calibration/validation (CAL/VAL) activities for Earth-observation missions. In this internship, you’ll explore how AIT captures and routine onboard imagery can support optical performance testing and radiometric calibration, and turn that work into an evaluated engineering prototype.

Location: Toulouse office, city center (Jeanne d’Arc metro station)
Duration: 6-month, end-of-studies internship— target start date: early 2027.

About this role:

• Review DSNU, radiometric calibration, and relevant EVP processing and CAL/VAL practices.
• Work with EVP engineers to identify suitable routine images, metadata, and reference calibration data; document data limitations and provenance.
• Prepare a reproducible dataset and establish reference baselines, including a holdout evaluation strategy that avoids tuning and testing on the same acquisitions.
• Implement a baseline estimator: a data-fit term combined with a regularizing prior.
• Investigate additional AI/ML methods only where suitable training or validation data are available; compare them against the baseline rather than assuming they are superior.
• Evaluate correction quality, residual fixed-pattern structure, stability across images, runtime, and failure cases.
• Package the workflow as a prototype script or notebook with plots, quality checks, and an auditable record of inputs, parameters, and outputs.
• Present findings and recommendations to EVP and CAL/VAL stakeholders; document what evidence would be needed before operational use.

The intern will gain experience in:

• Radiometric calibration and CAL/VAL for space-based optical imaging.
• Push-broom image characteristics and PRNU/DSNU correction.
• Applied optimization and, where appropriate, machine learning for image processing.
• Experimental design, reference-data limitations, and quantitative performance assessment.
• Building reproducible engineering tools and communicating technical evidence to mission teams.

Requirements

• Student in engineering, applied physics, remote sensing, image processing, data science, or a related field.
• Fundamentals of signal or image processing and a strong interest in Earth observation.
• Basic programming skills in Python and comfort working with scientific data.
• Analytical mindset, rigor, and ability to document assumptions and limitations.
• Interest in calibration, validation, sensor behavior, or image quality assessment.
• Exposure to satellite imagery or remote-sensing data.
• Experience with optimization, statistical estimation, or machine-learning evaluation.
• Familiarity with image-analysis libraries and scientific Python tools.
• Understanding of radiometric calibration or data-quality assessment.

Benefits

• Gratification: €1,300 gross/month.
• Lunch vouchers (tickets restaurant), because let’s be honest, we love food! (we even have a Slack channel about it: #loft-gourmand)
• Sustainable mobility package (forfait mobilité durable).
• A genuinely international environment comprising 20+ different nationalities.
• A genuine mentoring program to support you on a daily basis throughout the internship.