Saurabh Shukla

About Saurabh Shukla

At Cantina Labs, I spearhead training video and image generation AI models that power our interactive AI bots at the Cantina platform. In my previous role as the head of AI, I led the development of a generative AI-based game engine. The game engine used various generative AI technologies such as Stable Diffusion, Large Language Models, Dreambooth, IP-Adapter, Control-Net, LORA, and more. I led the pre-training, fine-tuning, and evaluation of the models and worked with the engineering team for the model deployment. Our game engine used many-layered tech stacks for features such as text-to-game asset creation, reskinning, and automated level generation. I also formulated and deployed an evaluation pipeline for all the features using human labeling and RLHF. I am also passionate about AI education for experts and beginners alike. Follow me if you want to learn about AI fundamentals and the latest AI research. Some of the technologies which I have implemented are (not an exhaustive list): Stable diffusion, SDXL, LLMs, LoRA, dream booth, ControlNet, IP-adapters, Machine learning, Deep learning, Reinforcement learning, Computer Vision, Convolutional and Generative Neural Networks (GANs), Encoder-decoder networks, Azure, AWS, Databricks, MLFlow, Optimization, multivariate analysis, PCA, PyTorch, TensorFlow, Keras, SciKit-Learn, Python, MATLAB, R

Experience

1. Ai Research Scientist

   Cantina Labs

   Sep 2024 — Present

   Making AI characters speak. Leading training of world’s fastest and robust talking video models to power Cantina app. Download today and experience a new age communication and content creation.

Education

• University of Illinois Urbana-Champaign

  Doctor of Philosophy - PhD

• Indian Institute of Technology, Delhi

  Bachelor of Technology (B.Tech.)