Research Scientist at Skyfall AI
Job Overview:
As an applied research scientist at Skyfall you will be responsible for advancing the state of the art for agentic systems. As part of your work you will develop new methods and techniques fo reasoning and planning with Large Language Models (LLMs), multiagent systems, and model based reinforcement learning to help reach Skyfall’s goal of developing a fully autonomous enterprise. We are seeking individuals who are passionate about areas such as program synthesis, natural language processing, conversational systems and reinforcement learning.
Research Key Responsibilities:
Develop new techniques and methods for advancing the state of the art in the areas of reinforcement learning with LLM , planning and reasoning with LLM and multi agent systems
Develop methods and techniques for advancing the state of the art for multi agent agentic systems
Conduct experiments and build datasets that improve LLM performance in complex, context-dependent workflow
Stay up to date on the latest advancements in academic research related to planning/reasoning with LLM , Reinforcement Learning and Multi Agent Systems
Qualification
PhD in Computer Science, Machine Learning, Artificial Intelligence, or a related field (preferred) or Masters in Computer Science , Machine Learning , Artificial Intelligence or related field and at least 5+ years of industry research experience
Experience with large models, a strong passion for the field, familiarity with large model technology development, and an understanding of the latest advancements in the domain.
Experience with large language model technologies such as Open AI GPT , Llama etc.
Experience with RL techniques such as Reinforce, PPO etc
Experience with neural network frameworks such as Pytorch , Tensorflow et
Preferred Skills & Experience:
Experience developing methods and techniques for planning with LLMs
Expertise in Reasoning with LLMs to address nuanced decision-making processes required in production models
Experience with multi agent reinforcement learning or multi agent planning
Preference for candidates with publications in well-known journals, blogs, or open-source projects in the fields of large language models, planning and reasoning with LLMs, multi-age reinforcement learning, or model based reinforcement learning
Location - Toronto
