LLM Algorithm Engineer

About Manus AI

Manus AI is an advanced general-purpose AI agent designed to seamlessly connect cognition with action, managing intricate workflows from inception to completion. It empowers users to delegate extensive tasks such as deep research, data analysis, and software engineering, transforming sophisticated AI technologies into dependable and practical outcomes.

Role Overview

We seek an LLM Algorithm Engineer to enhance large language model (LLM) and AI agent capabilities through rigorous post-training research and experimental development. This role involves working with diverse data strategies, training techniques, and evaluations to pinpoint task performance issues and convert research breakthroughs into product improvements.

Key Responsibilities

• Conduct research to improve LLM post-training techniques and training protocols, creating experiments focused on target tasks to boost performance and model generalization.
• Develop and refine training datasets by modulating quality, distribution, and application strategies to meet capability objectives and experimental feedback.
• Design and implement benchmarks and evaluation frameworks that assess key capabilities and practical scenarios, ensuring reliable comparisons of training methodologies.
• Examine model behaviors and recurring errors, generate hypotheses for improvements, and validate them through controlled training experiments, documenting findings and contexts comprehensively.
• Collaborate with product, engineering, and partner teams to integrate user and application feedback into data curation, training, and evaluation pipelines, and evaluate effectiveness on actual tasks.

Required Qualifications

• Strong foundational knowledge in machine learning and LLMs, with hands-on or research experience in large-scale foundation model training or post-training phases, including familiarity with supervised fine-tuning (SFT), reinforcement learning, or preference optimization.
• Expertise in leading deep learning frameworks, competent in independently setting up experiments, interpreting training outputs, and troubleshooting issues.
• Comprehensive understanding of the influence of data composition and training methodologies on model efficacy, with the ability to strategize enhancements based on evaluation results and practical applications.
• Active participation in relevant AI research and the capacity for independent research judgment, evidenced by formulation of testable hypotheses from open-ended challenges and thorough evaluation of outcomes.

Additional Information

Manus AI excels at automating diverse professional and personal tasks, achieving results efficiently to allow users to focus elsewhere.

What Manus AI Offers

• Opportunity to operate at the cutting edge of AI agent development, collaborating with a dynamic team to translate pioneering AI advancements into tangible impacts.
• Equity participation through a share incentive program, enabling employees to benefit from the company's long-term success.
• Unlimited access to Manus Tokens designed to encourage experimentation, creation, and boosted productivity.

If you are enthusiastic about innovative technology and aspire to make a meaningful contribution, we encourage you to explore this opportunity.