Constructor is an all-in-one platform for education and research. With expertise in machine intelligence and data science, Constructor is built to cater to the needs of schools, higher education, corporate training, alternative credentials, and professional sports, offering solutions for teaching and administration, learning and research.
From infrastructure to applications, Constructor elevates learning experiences, empowers educators, and drives research breakthroughs.
Our headquarter is situated in Switzerland. Also we have entities in Germany, Bulgaria, Serbia, Turkey, and Singapore.
We are seeking a highly motivated and talented student intern to join our dynamic team at Constructor Technology. As a student intern, you will have the opportunity to gain valuable hands-on experience in the EdTech industry while working alongside our experienced professionals.
In this role, you will assist our team in various software development, AI research or industry/market research and analytics projects. You will also have the chance to collaborate with cross-functional teams, contribute ideas, and learn from industry experts. This internship is an excellent opportunity for students to enhance their skills and kick-start their career in software development, AI research or market research.
This is a part-time opportunity (14 hours/week). Interns are not eligible for equipment.
Potential topics:
A. AI Research:
Text (speech) conditioned gestures and motion generation (for Avatar):
- Diffusion models for co-speech gestures generation.
- Diffusion models for text-to-motion generation.
- Complex gestures generation (like glassboard writings gestures).
- Generateplausible gestures for the specific person (few-shot adaptation).
- Reinforcement learning for speech (text) conditioned gestures generation planning.
- Text (speech)-to-motion multimodal foundation model (representation learning).
Skeleton-to-body generation with diffusion models (for Avatar):
- Few-shot skeleton-to-body model learning for the specific person.
- New gesture clip generation (conditioned on motion skeleton).
- Avatar clips generation with reinforcement learning (denoising diffusion policy optimization).
Advanced Analytics in LMS for Predictive Modelling
- Use advanced analytics from LMS data to predict student learning outcomes and develop intervention strategies for at-risk students.
- Model the relationship between engagement metrics in LMS and academic performance.
Automated Content Generation for Coding Education (LLM fine-tuning opportunity)
- Develop a system within the LMS for automated generation of coding exercises using machine learning, enhancing the Coding Lab experience.
- Evaluate the effectiveness of automated feedback and assessment in improving coding skills.
Reinforcement Learning for Adaptive Learning Pathways in LMS:
- Research the application of reinforcement learning to create adaptive learning pathways within the LMS, dynamically adjusting educational content based on student performance.
Predictive Analytics for Student Success in LMS:
- Apply ML techniques to predict student success and identify at-risk students by analyzing behavioral data within the LMS.
Diffusion Models for Creative Learning Content Generation (could be a tough one):
- Investigate the use of diffusion models for generating creative and engaging learning content, such as visual aids or interactive elements within the LMS.
Deep Learning for Avatar Emotion Synthesis:
- Investigate deep learning models that can synthesize natural-looking facial expressions for avatars based on the context of the educational content.
Requirements for AI projects:
- Currently enrolled as a student in CUB
- Strong analytical and problem-solving skills
- Python knowledge, incl. packages for ML (sklearn, tensorflow/pytorch)
- Data Science/AI algorithms knowledge
- Excellent problem-solving and communication skills
B. CS research topics
- “How to create Skynet» - distributed extensible computation environment with build-in fast tolerance
Requirements: knowledge of Rust/C/C++, a general idea of microkernels, and the operation of the OS, Golang knowledge is a plus
- Persistent OS in new environment - IoT, power consumption control and new code development model (C)
Requirement: knowledge of distributed algorithms, understanding the basic principles of discrete mathematics
- Container technology implementation for L4 microkernels with deep integration of native features via Genode OS framework
Requirements: knowledge of C/C++, a general idea of microkernels, and the operation of the OS, Golang knowledge is a plus
EdTech Industry Analytics:
- Market research of EdTech/ScienceTech industry
Requirements: good problem solving skills, critical thinking, ability to search, process and analyse vast amount of data about the market, PowerPoint/MS Word knowledge for reports preparation
What We Offer
