AIR Platform Unit

AIR Platform Unit is pledged to productize Megazone AIR Platforms (AIR Studio, AIR Data Hub, and AIR AIOps), a pivotal part of MegazoneCloud’s new AI brand called Megazone AIR.
People with diﬀerent skills are working together: AI/ML engineering, backend/frontend development, platform engineering, UI/UX design and publishing, and product management.

AI Engineering Unit

AI Engineering Unit aims at serving as an AI-tech powerhouse that firmly backs the company’s game plan for maximizing its presence in the domestic and global AI markets.
Our work is twofold: (1/application) Develop problem-solving playbooks that eﬀectively tackle customer headaches, and demonstrate the full significance and impact of leveraging Megazone AIR Platforms by creating concrete yet convincing illustrations of successful AI adoption. We do professional services strategically to achieve this goal. (2/foundation) Develop novel AI methods that answer the calls for optimizing the costs that arise throughout enterprise AI operations. The goal of our cutting-edge research is to diﬀerentiate the Megazone AIR product line and take it to the next level.

Develop a new AI method, called Brain AI, that automatically recommends cost-optimal execution environments for LLM- and LVM-based AI workloads in the multi-cloud setup (AWS, Google Cloud, and NHN Cloud).
Design Brain AI based on a Transformer-like architecture, and improve the architecture and the learning algorithm to achieve state-of-the-art performance.
Collect and analyze the execution data of AI workloads in the cloud, and train and validate Brain AI while fully exploiting our compute cluster built atop NVIDIA L40S GPUs.
Collaborate eﬀectively in making real progress at diﬀerent stages of research, from problem definition to data collection and analysis to method design to experiments and evaluation.
Conduct new research that will strengthen Brain AI in relevant directions while advancing the Megazone AIR product line by incorporating the research outcomes into it.

Experience publishing academic research papers as a lead author (peer-reviewed international conferences and journals).
Master’s degree in computer science, electrical engineering, and related fields OR equivalent experience.
Responsible and detail-oriented with strong self-motivation to produce exceptional work that presents both scientific novelties and practical impacts.
Ability to ask important questions and to formulate precise problem statements from fuzzy observations.
Ability to make constant progress in situations where a great degree of uncertainty and ambiguity prevails.
Capable of exercising hands-on leadership (coding, data collection, experiments, etc.) along with a sound grasp of concepts and theories.
Honest about your knowledge and open to constructive discussions, especially when things get diﬃcult.

Publication record at top-tier machine learning conferences (NeurIPS, ICML, ICLR, CVPR, etc.).
Ph.D. degree in computer science, electrical engineering, and related fields.
Thorough understanding of deep learning and recent advances in LLMs.

Present your research work (one or two) with crystal clarity while ensuring the following or comparable flow: motivation → problem → hypothesis → experiments → results → limitations and/or future work.
Share the moral of your experience with special focus on the insights you gained.