Qualifications
Key ResponsibilitiesIndependently design and execute experiments utilizing LLMs, prompt engineering, and innovative strategies to elicit and aggregate forecast relationships.Rapidly build MVPs and prototypes for testing and iteration of ideas.Devise and implement methods for eliciting, aggregating, and visualizing coherence rules across forecasts, including causal and correlational relationships, martingale processes, and necessary/sufficient conditions.Create user-friendly interfaces that render complex forecast relationships and world models comprehensible to stakeholders.Develop AI-enhanced tools for world model elicitation and clarification, including autonomous coherence rule detection and suggestions for latent variables.Integrate novel functionalities into the primary Metaculus platform and potentially create standalone tools for specialized applications.Technical RequirementsFrontend: Proficiency in JavaScript, TypeScript, React, Next.js, D3, Tailwind CSS.Backend: Experience with Python, Django, PostgreSQL, Redis, Sentry, Heroku, AWS.Collaboration: Familiarity with GitHub, Slack, Notion, Figma.Essential QualificationsClean Code Practices: A strong commitment to writing maintainable, testable, and well-documented code.Problem Solving: Excellent analytical skills with the capability to address ambiguous challenges in novel domains.AI Integration: Experience with prompting, retrieval-augmented generation (RAG), fine-tuning, etc.Communication: Effective communication skills to convey complex ideas clearly.
About the job
Metaculus stands at the forefront of public forecasting, boasting more than 2,800,000 forecasts spanning over 20,000 questions. We function as an essential epistemic infrastructure akin to Wikipedia and Our World In Data, dedicated to delivering reliable and precise information across a spectrum of topics, particularly within the realms of science and technology. Our mission is to empower individuals to model, comprehend, predict, and effectively navigate the world’s most significant and intricate challenges.
As a Research Engineer with a specialization in Forecast Coherence & World Models, you will be a trailblazer in crafting technologies that enhance the transparency and accessibility of forecasts and their foundational world models. This role positions you at the dynamic intersection of software engineering and epistemics, where you will develop tools that assist domain experts and forecasters in creating more coherent predictions, thereby facilitating improved decision-making on multifaceted, interdisciplinary issues.
Your responsibilities will involve a strong emphasis on independent experimentation and rapid prototyping. You will conduct experiments using large language models (LLMs), design and test innovative methodologies, and create minimum viable products (MVPs) to validate conceptual ideas. Additionally, you will devise methods to elicit, aggregate, and visualize the interconnections among forecasts, including causal relationships and necessary conditions, while incorporating AI to aid in constructing world models and identifying critical variables that could enhance forecast accuracy.
About Metaculus
At Metaculus, we are dedicated to providing a platform that not only aggregates forecasts but also empowers users with the tools necessary to engage with complex predictive analytics. Our innovative approach to epistemic infrastructure aims to transform how individuals and organizations approach forecasting, ultimately leading to better-informed decisions in critical areas such as science and technology.
