Publisert:
28.10.25
Foundational Model Research & Engineering Internship: Building Next-Generation Industrial AI
About Cognite
Embark on a transformative journey with Cognite, a global SaaS forerunner in leveraging AI and data to unravel complex business challenges through our cutting-edge offerings including Cognite Atlas AI, an industrial agent workbench, and the Cognite Data Fusion (CDF) platform. We were awarded the 2022 Technology Innovation Leader for Global Digital Industrial Platforms & Cognite was recognized as 2024 Microsoft Energy and Resources Partner of the Year. In the realm of industrial digital transformation, we stand at the forefront, reshaping the future of Oil & Gas, Chemicals, Pharma and other Manufacturing and Energy sectors. Join us in this venture where AI and data meet ingenuity, and together, we forge the path to a smarter, more connected industrial future.
Our values
Impact: Cogniters strive to make an impact in all that they do. We are result-oriented, always asking ourselves.
Ownership: Cogniters embrace a culture of ownership. We go beyond our comfort zones to contribute to the greater good, fostering inclusivity and sharing responsibilities for challenges and success.
Relentless: Cogniters are relentless in their pursuit of innovation. We are determined and deliverable (never ruthless or reckless), facing challenges head-on and viewing setbacks as opportunities for growth.
The Cognite Atlas AI research team is engaged in world leading work on foundational transformer models for industrial use cases, pushing the boundaries of applied artificial intelligence. While traditional Large Language Models (LLMs) excel at general tasks, they fall short on some critical industrial problems. This is because their training data, sourced primarily from the public internet, lacks the specialized, proprietary knowledge required to understand complex industrial processes, equipment, and data. Cognite has access to this type of data that almost no other product company in the world does.
Our mission is to bridge this gap by developing and scaling the next generation of AI systems trained on domain-specific industrial data. This initiative involves fundamental research into model architecture and pioneering ML engineering to build models that can reason over and understand complex industrial data.
This internship project offers a unique opportunity for exceptional students passionate about Deep Learning and ML Engineering to contribute directly to the core of our AI strategy. The intern will be immersed in the end-to-end lifecycle of foundational model development, from experimenting with novel architectures to building the high-performance infrastructure required to train them. This role is for those who want to build the future of industrial AI from the ground up.
This internship will span 6-8 weeks, commencing in the first week of July. Interns will work collaboratively in pairs of two, fostering a dynamic and supportive learning environment. For students with exceptional outcomes during the internship, continued collaboration is likely to be offered.
Project Scope & Activities
- Foundational Model Research & Development:
- Implementing and experimenting with novel transformer architectures, attention mechanisms, and optimization strategies for industrial data.
- Contributing to the pre-training and validation of large-scale models on massive, proprietary industrial datasets.
- Conducting rigorous analysis and ablation studies to understand model behavior and drive architectural improvements.
- ML Engineering & Scaling:
- Designing and building robust, scalable data pipelines for large-scale model training.
- Developing and maintaining the infrastructure for distributed training across cutting-edge cloud and on-premise GPU clusters.
- Creating sophisticated tools and frameworks for model evaluation, performance profiling, and experiment tracking to accelerate the research and development cycle.
Expected Outcomes
- Contributed directly to the development of a state-of-the-art foundational model tailored for industrial applications.
- Gained significant hands-on experience in large-scale, distributed model training and high-performance computing.
- Designed and implemented critical components of an MLOps pipeline for foundational model research.
- Acquired a deep, practical understanding of the challenges and solutions in training and scaling massive AI models.
- Authored clear technical documentation on model experiments, architectural designs, and engineering systems.
Required Skills & Qualifications
- Machine Learning / AI: Deep theoretical and practical understanding of deep learning, particularly transformer architectures.
- Python: Advanced proficiency in Python and deep experience with at least one major ML framework (e.g., PyTorch, TensorFlow, JAX).
- Software Engineering: Strong fundamentals in data structures, algorithms, and software design principles.
- Problem-Solving: Excellent analytical and problem-solving skills, with a passion for tackling complex, open-ended challenges.
- Collaboration: Ability to work effectively in a team, communicate technical concepts clearly, and adapt to a fast-paced research environment.
Bonus Skills (Nice to Have):
- Experience with cloud platforms (GCP, AWS, Azure) and distributed computing.
- Familiarity with containerization and orchestration technologies (Docker, Kubernetes).
- Experience with high-performance computing (HPC) environments and GPU optimization.
- Contributions to open-source ML or systems software projects.
Cognite
The key to industrial digitalization lies in data liberation. Heavy-asset industries like oil and gas, shipping, manufacturing, and power and utilities already have the data. Now they need software to collect, clean, and contextualize the data. A resource to transform the data into information and to stimulate a thriving ecosystem of industrial applications.
Cognite Data Fusion (CDF) presents a digital representation of industrial reality to make it accessible and meaningful for humans and machines.
With CDF, our industrial customers can harness the potential of advanced analytics, deploy algorithms, and build customized applications. We make it possible to maximize the strategic value of data. Realizing the promise of digitalization
To succeed, we need a lot of skill-sets, such as backend programming with large scale distributed systems, real-time systems, machine learning, optimization, web frontends, 3D-models, robots and more. We need project managers who can be consultants for our customers. It will be a very exciting environment where team members will learn new skills from some of the best.
Why work for Cognite?
- You will have a real impact on our customers and Cognite
- Free snacks and drinks throughout the day
- Opportunity to work for, and contribute to the growth of one of the most exciting and fastest-growing new software companies in the world
- Competitive salary and benefits (including pension plans, insurance, parental benefits and more)
- Coverage of mobile telephone subscription and broadband connection
- Extended private health services and free yearly health check
- Subsidized lunch at the canteen, with various food options (pizza/sushi)
- Free staffed gym
- Social activities (book club, team sports activities - football, boxing, regular Cognite social events)
- Free online Norwegian courses for levels A1 and A2
