AI Engineer (Madrid)

Duckbill is revolutionizing the personal assistant market by harnessing the power of AI to offer seamless, personalized solutions for daily life management. This innovative service is dedicated to simplifying complex tasks, from planning trips to managing appointments, through intelligent, user-centric technology.

The AI Engineer role at Duckbill is a cornerstone of this ambition, aimed at a highly skilled professional who is both proficient in current AI technologies and eager to push the boundaries of what's possible. This position is central to Duckbill's commitment to innovation and operational efficiency, requiring a unique blend of technical mastery, creativity, and business insight. AI engineers will lead the development of sophisticated AI solutions that not only enhance user experience but also drive the company's strategic objectives, setting new standards for personalized assistance.

What you'll do

• Design groundbreaking AI solutions and develop prototypes that drive the future of Duckbill's user-facing experience.
• Leverage the latest innovations in AI and ML to provide a hyper-personalized experience to our users, including the development of high-engagement recommendation systems.
• Build tools to maximize the output and quality of our human personal assistants (copilots), reimagining the collaboration between humans and machines in the virtual assistant space.
• Automate task execution by leveraging the entire AI stack, from traditional ML to the latest LLM-agent frameworks.
• Improve human task execution quality by enhancing our intelligent information retrieval systems, utilizing search and RAG techniques.
• Contribute to our operations team's excellence by building smart monitoring and intervention systems.
• Enhance our AI ops stack by designing and improving our evaluation frameworks, allowing the team to iterate on prompts, models, and approaches with both agility and thoroughness.
• Enhance our human annotation flow stack to enable leveraging our human resources for annotation, model comparison, and golden dataset generation.
• Maximize the value extracted from our proprietary data using state-of-the-art techniques (fine-tuning, RLHF, RAG, etc.), critically contributing to the defensibility of our AI strategy.
  

What we look for

• 4+ years of experience in an ML engineering role.
• Proficiency in machine learning, with experience in areas such as Generalized Linear Models, Gradient Boosting, Deep Learning, and Probabilistic Modeling. Specialized knowledge in recommender systems is a plus.
• Strong engineering skills in Python and data manipulation skills in SQL.
• Solid knowledge of the main patterns in ML model deployment, both online (API-based) and offline (batch jobs), with the ability to design production-grade solutions.
• Experience with the entire MLOps stack, especially in experiment/model performance tracking and model monitoring in production.
• Strong product and business intuition, with a deep understanding of what makes an ML project a business success.
• Knowledge of the foundations of the LLM stack, including main foundational models (both closed and open-source), prompt engineering techniques, evaluation frameworks, and function calling.

Nice to have

• Strong understanding of embeddings and RAG systems.
• Basic knowledge of LLM agent architectures and the main stack libraries (e.g., langchain, llama_index).
• Prior experience with any of the components of our backend stack, such as Django, Celery, MongoDB, and PostgreSQL.

What we offer

• Competitive base salary plus stock option package.
• International, positive, dynamic, and motivated work environment.
• 10% of the time dedicated to off-backlog research on applying SOTA to Duckbill’s challenges.
• Hybrid work model (highly flexible, with a preference for two in-person days per week).
• Health insurance.