Lead, AI Trust & Safety Research Engineer /Scientist

About the Company: One of our well-established Start-up clients in Bengaluru, India are seeking an experienced Lead AI Trust and Safety Research Scientist to spearhead our efforts in developing safe, ethical, and reliable AI systems. In this role, you will work closely with our product and research teams to ensure our foundation models, including large language models and multimodal systems, adhere to the highest standards of safety, honesty, and helpfulness. Your work will be crucial in shaping the responsible development and deployment of AI technologies that can positively impact millions of lives across India and beyond.

Responsibilities:

• Collaborate with the product team to define and implement safety aspects that ensure our AI models deliver helpful, honest, and harmless outputs.
• Lead the development and implementation of advanced safety techniques for large language and multimodal models, including but not limited to Reinforcement Learning from Human Feedback (RLHF), Direct Preference Optimization (DPO), and Proximal Policy Optimization (PPO).
• Conduct thorough red teaming exercises to identify potential vulnerabilities and ensure our models do not generate harmful or undesirable high-quality outputs.
• Design and oversee the generation of high-quality synthetic data for fine-tuning models to improve safety and performance.
• Lead research initiatives in AI safety, focusing on models with 10s and 100s of billions of parameters.
• Develop innovative methodologies for evaluating and mitigating risks associated with large-scale AI models.
• Stay abreast of the latest developments in AI ethics and safety, and contribute to the broader scientific community through publications and conference presentations.
• Mentor and guide junior researchers in AI safety principles and practices.
• Collaborate with cross-functional teams to integrate safety measures throughout the AI development pipeline.

Qualifications:

• M.S. or Ph.D. in Computer Science, Machine Learning, or a related field with a focus on AI safety and ethics.
• 4 years of experience in AI research, with a strong emphasis on trust and safety aspects of large language models and multimodal systems.
• Demonstrated expertise in implementing advanced safety techniques such as RLHF, DPO, and PPO for foundation models.
• Proven track record of building or fine-tuning models with 10s and 100s of billions of parameters.
• Strong background in red teaming methodologies and adversarial testing for AI systems.
• Experience in generating and working with high-quality synthetic data for model fine-tuning.
• In-depth understanding of ethical considerations in AI development and deployment.
• Excellent programming skills, particularly in Python and deep learning frameworks such as PyTorch and TensorFlow.
• Strong publication record in top-tier AI conferences and journals, specifically in areas related to AI safety and ethics.
• Exceptional communication skills with the ability to explain complex technical concepts to both technical and non-technical audiences.
• Experience leading research teams and mentoring junior researchers.
• Familiarity with relevant regulatory frameworks and industry standards for AI ethics and safety.

Please coordinate with Jia for more information and setup sometime for a quick chat to discuss the role and client.