• Diseñar y desarrollar modelos de machine learning e IA para casos de negocio. • Entrenar, evaluar y desplegar modelos utilizando Amazon SageMaker. • Implementar soluciones de IA generativa con Amazon Bedrock. • Preparar y transformar datos asegurando calidad y disponibilidad. • Colaborar con equipos para llevar modelos a producción bajo buenas prácticas de MLOps.

• Rapidly stitch together estimates of relevant variables from data you can rapidly obtain (e.g. health harm caused by different Climate TRACE assets) • Iteratively improve estimates by continuously improving algorithms with new input data using Bayesian updating or similar methods • Pull, clean, format, and organize relevant data from the web, satellites, our databases, and partners • Write clear methodologies, API documentation, and papers for technical audiences • Write brief punchlines or blog posts for team members, policymakers, nonprofits, and corporate decision makers • Estimate the causal impact of various interventions (like building solar) • Quantify opportunities to increase impact by adjusting control variables (e.g. moving air quality sensors to where it will minimize GHGs) • Conduct formal analysis of the comparative shadow value of relaxing different control variable constraints across many complex causal chains • Apply this to inform the team’s data assimilation work: where would improving our data or methods cause the most impact?

• Develop technical solutions that meet business requirements, aiming to create scalable and high-performance applications; • Develop batch systems and REST APIs in Python based on provided documentation, creating applications that meet user requirements; • Maintain web portals; • Identify and fix incidents and maintain systems according to problem descriptions and log analysis; • Prepare technical documentation for implemented solutions; • Write unit tests; • Interact with the business area to clarify business requirements, validate developed systems, and provide technical input in alignment meetings; • Experience with Agile methodologies is desirable; • Develop and evaluate generative AI (LLMs) solutions applied to real business problems such as product recommendation, document analysis, chatbots, and collections strategies; • Conduct experiments to optimize models, including prompt engineering, fine-tuning, and RAG (Retrieval-Augmented Generation) techniques; • Define and implement evaluation metrics for generative models (e.g., quality, factuality, consistency, and bias); • Perform robust statistical analyses to validate results, including hypothesis testing, confidence intervals, and simulations; • Identify risks and propose solutions to mitigate hallucinations, biases, and reliability issues.

• Identify long-term research goals as well as intermediate milestones for AI models and applications. • Execute research that pushes forward the state of the art in AI domains such as predictive health outcomes, LLMs, and Generative AI. • Directly contribute to experiments training AI models, including experimental design, writing reusable code, running evaluations, and preparing results for broader circulation. • Publish research in top tier conferences and journals in the diabetes, health sensing, and technology communities. • Contribute to the evaluation and improvement of the safety of our models and systems, identifying areas of risk and proposing mitigation strategies.