Lead AI Engineer

About Avatar.One

Avatar.One is at the forefront of AI innovation, developing cutting-edge solutions that redefine the boundaries of conversational AI. Our platform is a web-based immersive environment for creating, chatting, and interacting with AI characters, designed to offer users uniquely engaging and realistic conversational experiences. We are passionate about pushing the limits of AI technology to create interactive characters that feel alive. Avatar.One is currently seeking a Lead AI Engineer to spearhead our efforts in enhancing the quality and capabilities of our AI models, ensuring our users experience the most natural and engaging conversations possible.

Role Summary

As a Lead AI Engineer at Avatar.One, you will play a pivotal role in identifying, benchmarking, and implementing superior AI models to elevate our conversational AI platform. You will lead projects aimed at enhancing the scalability, accuracy, and realism of our AI-driven characters, and improve the automation and emotional intelligence of our AI responses. Additionally, you will be responsible for optimising and retraining our LLMs with unique datasets to enhance the human-like quality of the conversational experience. This position offers the opportunity to work on groundbreaking AI projects and significantly impact the quality of our user experience.

Key Responsibilities:

• Conversational quality optimisations
  • Identify, evaluate and validate superior AI models to enhance conversational experiences for different use case like romantic relationships, coaching, education and roleplay. 
  • Implement and maintain a test environment for consistent performance evaluation.
  • Collaborate with team members to define benchmarking questions and approaches, ensuring the model's integration is seamless and efficient.
  • Test the product yourself regularly and review user feedback together with team to identify ways to improve the conversation quality. 
• Prompt engineering
  • Refine and optimise context prompt data structures to enable user customisation
    
  • Create controls and validation to ensure the quality of AI conversation across characters, both internal created and user generated
• Enable character scaling
  • Enable the scaling of character interactions on Avatar.One, ensuring compatibility and performance.
  • Validate and repurpose character data sets for enhancing LLM prompting techniques.
  • Develop automation pipelines for dataflows to facilitate easier integration and scalability.
• Automation and emotion classification development
  • Enhance AI response automation by improving emotion classification accuracy and the decision-making process for memory retention.
  • Address challenges related to automated responses and animation diversity by optimizing the call structure to the LLM.
  • Innovate solutions to ensure more human-like, dynamic and varied emotional expressions, contributing to a more realistic and engaging user experience.
• LLM optimisations
  • Lead efforts to retrain Large Language Models (LLMs) using unique datasets to significantly improve the human-like quality of the conversational experience for a range of use cases. 
  • Help identify and validate unique data sets for training models. 
  • Oversee the retraining process, ensuring the integration of new data enhances the AI’s understanding and response accuracy while maintaining ethical AI practices.

• Bachelor’s degree in Computer Science, AI, Machine Learning, or a related field. A Master’s degree or higher is preferred.
• Minimum of 1 year of hands-on experience with Large Language Models (LLMs) and AI conversational platforms.
• Proven track record in AI model benchmarking, NLP, and machine learning algorithms.
• Experience with character AI and animation integration is highly desirable.
• Strong programming skills in Python, including experience with AI and machine learning libraries.
• Ability to work independently in a remote setting, with strong communication skills and a collaborative team spirit

Please mention the word **LIKABLE** and tag RNDQuMjM0LjE4Ni4xNTg= when applying to show you read the job post completely (#RNDQuMjM0LjE4Ni4xNTg=). This is a beta feature to avoid spam applicants. Companies can search these words to find applicants that read this and see they're human.