Applied AI Scientist

As an Applied AI Researcher in the deep-tech and AI research industry, you will be an integral part of our advanced research team. Your primary responsibility will be to design innovative multi-agent systems, develop domain-specific scaffolding techniques, and construct evaluation frameworks for cutting-edge AI systems. By combining LLMs, reinforcement learning, and multi-agent systems, you will play a crucial role in creating scalable and impactful solutions at the forefront of applied AI. Your key responsibilities will include: - Architecting and implementing novel multi-agent systems to facilitate advanced problem-solving and collaboration - Designing domain-specific scaffolding techniques to customize AI behavior for complex, real-world domains - Curating and managing high-quality datasets for training and evaluating AI models across scientific and industrial applications - Establishing robust evaluation frameworks to assess performance, alignment, and robustness - Researching and applying reinforcement learning techniques such as RLHF, DPO, and GRPO - Exploring post-training optimization, fine-tuning, and domain adaptation methods - Collaborating cross-functionally with engineering and product teams to transition research into production-ready solutions - Staying informed about the latest developments in AI and actively contributing to internal and external research communities - Documenting and communicating findings through technical reports, presentations, and publications To be successful in this role, you should possess the following qualifications: - A Masters or PhD in Computer Science, Artificial Intelligence, Machine Learning, or a related field - 4+ years of experience in applied AI research or equivalent industry R&D experience - Strong foundations in optimization, probability, and linear algebra - Expertise in Python and frameworks like PyTorch or JAX - Experience with RL and post-training methods (e.g., SFT, DPO, RLHF) - Proficiency in building and aligning small language models (SLMs) and reasoning-specific models - Familiarity with prompting strategies like Chain of Thought and dataset design for reasoning tasks - Deep understanding of multi-agent systems and distributed training (multi-GPU/multi-node) - Experience in designing evaluation metrics and performance analysis methodologies Preferred experience includes: - Publications in leading ML conferences (NeurIPS, ICLR, ICML, AAMAS, etc.) - Applying AI to scientific domains such as drug discovery, chip design, or materials science - Exposure to multimodal models and VLMs - Experience with open-ended systems and emergent behavior in agent-based learning - Background in computational science (chemistry, physics, EE, or applied math) - Familiarity with MLOps, Kubernetes, Ray, Hydra, and MLflow - Experience with domain adaptation, interpretability, and model optimization for deployment - Contributions to open-source AI projects - Expertise in building GPU-accelerated pipelines and optimizing inference at scale In return, we offer you the opportunity to work on high-impact, frontier research with real-world applications, access to high-performance computing resources, a collaborative and intellectually stimulating environment, autonomy to explore novel ideas aligned with our mission, and a competitive salary, benefits, and opportunities for growth.,