Lead Research Scientist - Generative AI

With over 50 products and 1.5 billion installs worldwide, AIBY is at the forefront of mobile-first product development with a focus on AI technology. Our mission is to make artificial intelligence accessible to people from all walks of life. Our open-minded approach to ideation sessions, in which every member of the team can make their voice heard, innovation, and team spirit are the backbones of our company. Thanks to them, we are proud of titles like ChatOn, Arta, Forma, Plantum, Fonts Art, and Themify.

To further expand our horizons and achieve even bigger results in a number of fields, we’re looking for a Lead Research Scientist to join our team, who will stay updated with the latest industry trends, tools, and techniques in Generative AI and apply them to enhance our AI capabilities.

What you will do:

- Optimize deep neural networks for fast inference (distillation, quantization, research of new architectures, etc.)

- Train and fine-tune multi-modal diffusion models using custom datasets

- Deploy generative models using various inference engines like TensorRT, AITemplates, and CoreML

- Research and experiment to improve existing models and propose new approaches

- Work with the Engineering Team, visualize and report the research results to the Product Team.

Tech stack:

Inference Engines & deploy

- TensorRT + NVidia Triton Server + FastAPI

- AITemplate + Lightning (FastAPI) + RabbitMQ

- CoreML – will be a plus

ML frameworks:

- PyTorch + JAX

- Huggingface: diffusers, transformers, tokenizers.

ML/DL approaches:

- Distillation, quantization

- Diffusion Models

- Generative Adversarial Networks (adversarial training, perceptual loss)

- Super Resolution

- Traditional Computer Vision: object detection, classification, semantic segmentation (will be a plus).

What you should have:

- Python and PyTorch proficiency

- Hands-on experience with the development process in DL and ML (Docker, git, TensorBoard, DVC, Weights & Biases)

- Experience with a similar stack to ours

- Understanding of modern computer vision concepts like diffusion models, GANs, and LLMs

- At least a B1 level of English.

Will be a plus:

- Public activity on Kaggle, GitHub, Leetcode, or other media

- Software development background

- At least a B1 level of Belarussian/Ukrainian/Russian.

- A PhD in Computer Vision or a relevant technical field

Why work with us?

- Competitive salary with a generous bonus

- Variety of unique projects

- Independent decision-making and no meaningless paperwork

- Opportunity to directly influence the development of our products

- Opportunity to work from anywhere

- Sports reimbursement program

- Health Benefits (depends on the location)

- Reimbursement for professional courses and training

- 21 working days of paid vacation

- 5 sick days and other benefits

‍