- Design, prototype, and refine deep learning models for text-to-video and image-to-video generation using generative AI (e. g., diffusion models, transformers, temporal GANs).

- Work with the lead data scientist on ongoing challenges to define use cases, set priorities, and bring production-grade AI systems to life.

- Develop robust pipelines for dataset curation, preprocessing, training, and model evaluation, including automated visual, temporal, and semantic quality metrics.

- Stay ahead of the curve in the fast-moving GenAI space; guide strategy, share insights internally, and represent the team in technical forums as needed.

- Take ownership of the work involved and be proactive in any communication.

- Involved in experimentation with SOTA models to achieve greater outcomes for JHS as a platform.

Requirements :

- Deep expertise in Transformer-based architectures (e. g., BERT, GPT, T5 ViT). The ideal candidate has a strong theoretical understanding and hands-on experience building, training, optimising, and deploying transformer models for real-world applications.

- Strong understanding and hands-on experience with LLMs, including fine-tuning, prompt engineering, and evaluating foundation models for multimodal generative applications.

- Proven experience with deep learning architectures for generative tasks, including diffusion models, transformers, and GANs.

- Solid grasp of computer vision, natural language processing, and multimodal learning concepts.

- Experience in training large-scale models, transfer learning, and fine-tuning custom datasets.

- Proficiency in Python, SQL, and deep learning frameworks such as TensorFlow and PyTorch.

- Strong first-principles thinking to translate business challenges into solvable scientific problems.

- Effective communication skills to articulate complex technical details and collaborate across teams.

- Proactive ownership and a problem-solving mindset to independently drive tasks forward.