Research Scientist / Research Engineer

📋 Description

• • Adaption Inc. is at the forefront of a paradigm shift in artificial intelligence, moving beyond static models to create systems that learn and evolve in real-time. Our core mission is to develop efficient intelligence that dynamically adapts to the world, making AI flexible, personalized, and universally accessible. We champion efficiency as the key enabler for expanding access and ensuring that groundbreaking innovations benefit a broad audience, not just a select few. Our commitment to talent density means we assemble a team of the most skilled and driven individuals dedicated to pushing the boundaries of continual adaptation. We are actively seeking innovative builders and creative thinkers who are eager to define the next generation of intelligent systems.
• • Our research is guided by a set of fundamental principles: meticulous attention to detail, a conviction-driven approach to problem-solving, and a focus on metrics that truly reflect real-world impact. We believe that technical excellence is achieved through an obsessive focus on every aspect of the system, co-designing serving infrastructure, algorithms, and user interfaces as an integrated whole to maximize efficiency and facilitate real-time adaptation. We operate with a sense of urgency as a cohesive team, strategically concentrating our efforts on high-conviction initiatives where research directly intersects with tangible impact. We are committed to measuring what matters, understanding that the most intelligent systems will be those capable of interacting meaningfully with the world. Research ideas are rigorously tested through functional products, and we critically evaluate whether an innovation truly matters if it doesn't demonstrably improve user capabilities. We prioritize sharing advancements that move the field forward, rather than optimizing solely for publication counts.
• • This research-focused role is intrinsically tied to achieving real-world impact. Our frontier innovations are centered around key areas: efficiency, gradient-free exploration, real-time learning, and interface design. We are witnessing a revolution where the declining cost of synthetic data generation allows us to treat the data space as a malleable resource that can be actively optimized. This enables us to steer synthetic data towards desired properties, making previously obscured or data-scarce domains more visible and accessible. The ultimate measure of an intelligent system will increasingly be its ability to interact effectively with the world. Consequently, researchers passionate about intelligence must now also be deeply concerned with how models engage and interact. If these principles and challenges resonate with your research aspirations, we strongly encourage you to connect with us.
• • Key responsibilities in this role include driving innovation by co-designing algorithms that respond in real-time to product signals and user feedback, and pioneering novel feedback mechanisms that enhance algorithmic performance. You will be instrumental in cross-stack optimization efforts, collaborating across software, hardware, and algorithmic domains to achieve system-wide efficiency improvements. A critical aspect of this role involves measuring what matters, aligning research efforts with demonstrable real-world impact, as our current algorithms are designed for direct interaction with the world, underscoring the paramount importance of product integration and performance.
• • We are looking for individuals who possess a PhD or equivalent research experience in a computer science discipline, coupled with 4-5+ years of relevant industry experience. Deep expertise in at least one of the following areas is essential: model efficiency, real-time alignment, or algorithmic optimization. A strong systems-thinking ability is required to understand and optimize the entire machine learning stack. Proficiency in Python programming and experience with deep learning frameworks such as PyTorch, JAX, or TensorFlow are crucial. Familiarity with model optimization techniques like Reinforcement Learning from Human Feedback (RLHF) and fine-tuning is expected. Prior experience within an industrial research lab setting, involving large-scale computing, is highly valued. Beyond technical qualifications, we seek exceptional teammates who bring a positive energy to their work, are unafraid to propose bold ideas, and demonstrate adaptability. We encourage applications even if not every qualification is met, as we value potential and a willingness to learn and grow.