Research Intern – Multimodal Computer Vision (RGB–Thermal Fusion)

**Company Description** At Keenfinity, we are a globally leading provider of **innovative and professional** security and communication solutions. With over 4,200 employees in over 50 countries worldwide, our ambition is clear: we offer more than just technology – we secure, connect, and amplify the moments that matter in life. Next to our passion for technology we’re very passionate about our work environment. Based on values such as trust, appreciation, and accountability we **all** work together to shape the future – boldly, customer-focused and with a strong team spirit. **Job Description** As an intern, you will work on advanced **multi-modal computer vision** problems, focusing on improving visibility and robustness in low-light environments by combining RGB and thermal imaging. * Get hands-on with data – Work with visible (RGB) and thermal image/video datasets, including internal and public sources * Build strong foundations – Understand and reproduce baseline pipelines for image registration and fusion * Define how success is measured – Design and implement a robust evaluation framework, including both quantitative metrics and visual assessment * Develop fusion algorithms – Implement and benchmark baseline RGB–thermal fusion methods * Push innovation with AI – Design and train a reliability-aware fusion model using techniques such as attention or gated fusion * Handle real-world challenges – Improve robustness for imperfect alignment and low-visibility scenarios using signals like alignment confidence and illumination conditions * Detect and reduce artifacts – Develop methods to identify and minimize issues such as ghosting, hallucinations, and image distortions * Compare and validate results – Evaluate your approach against baselines using performance metrics and qualitative insights * Explore real-world impact (optional) – Validate improvements on downstream tasks such as object detection in low-light conditions * Communicate your findings – Document your work in a thesis/report and prepare a publication-ready results package, including ablations and robustness analyses **Qualifications** ✅ Your Profile * You have experience with Python and deep learning frameworks (e.g. PyTorch, TensorFlow, or Keras) * You have a strong foundation in computer vision and deep learning * You are fluent in English, both written and spoken ⭐ Nice to have * Experience with multi-modal learning, image fusion, or low-light vision * Familiarity with image registration, calibration, optical flow, or feature matching * Experience with experiment design, including ablation studies, evaluation protocols, and reproducibility best practices * Interest in solving real-world challenges in AI for vision systems (e.g. surveillance, perception) **Additional Information** **💼 What We Offer:** * A hands-on internship in an international and collaborative team. * Exposure to diverse projects, stakeholders, and global ways of working. * Close guidance and mentorship from experienced professionals. * A dynamic role where your input, ideas, and initiative are truly valued. * The option to work from home one day per week. * An internship allowance between €500 and €600 per month (based on full-time employment). **📝 Application Process** * Interested? Hit the apply button! * Phone meeting with a Keenfinity campus recruiter; * Interview via Microsoft Teams with team members; * Contract offers from Keenfinity. **🌱 Join us at KEENFINITY!** At Keenfinity we don’t just build innovative solutions — we shape a smarter, more connected world through technology. We value different backgrounds, ideas, and experiences and we’re committed to growing, learning, and celebrating success as one team. Everyone is welcome here — we foster an environment where everyone is respected, valued, and encouraged to be their authentic self. Keenfinity is an equal opportunity employer, offering equal opportunities for all. We welcome applications from people with disabilities and can offer support, if needed. When everyone has a chance to contribute, we all do better.