About Mappa Mappa builds real-time voice-AI infrastructure that decodes over 100,000 behavior patterns per person—based on prosodic and linguistic signals—and turns them into actionable insights for hiring and beyond. Our data science team transforms raw audio into structured understanding at scale. Role Purpose Analyze and model complex behavioral and audio data sets to power Mappa’s AI systems. You’ll design and implement data pipelines, extract actionable insights, and prepare datasets for training cutting-edge machine learning models. Key Outcomes (First 6 Months) Deliver at least 3 key insights from exploratory analysis that inform product or model design. Build and deploy at least one production-grade ML-ready dataset using Airflow. Identify and implement ≥ 2 performance improvements in existing data pipelines. Collaborate with ML and Product to define data strategies for new features or behavior signals. Core Responsibilities Analyze voice, text, and behavioral datasets using statistical and ML methods. Prepare, clean, and transform data for downstream machine learning tasks. Design and maintain data pipelines using Airflow and Dockerized workflows. Contribute to model development using tools like Hugging Face and PyTorch. Collaborate with ML Engineers and Product to align data initiatives with product goals. Present findings and prototypes to cross-functional stakeholders, including CTO and COO. Monitor data health, quality, and coverage across all stages of the pipeline. Must-Have Skills 3–5 years in data science, ML, or analytics with experience in real-world data problems. Proficiency in Python and ML/data libraries (e.g., pandas, scikit-learn, Hugging Face). Experience building and maintaining data pipelines with Apache Airflow. Familiarity with Docker and modern data architecture principles. Strong foundation in statistics, data wrangling, and model evaluation. Comfortable communicating complex results to non-technical stakeholders. Fluent in English. Nice-to-Have Experience with audio, voice, or behavioral datasets. Familiarity with time-series analysis or sequence modeling. Knowledge of vector embeddings or NLP model fine-tuning. Who You Are A curious and rigorous thinker who’s fluent in data and comfortable navigating ambiguity. You love turning messy signals into structured insight—and insights into products. You value reproducibility, impact, and collaborating with multidisciplinary teams. Why Join Mappa Deep impact: Your insights shape the core of our AI product. Modern stack: Work with Python, Hugging Face, Airflow, and Docker in a fast-moving team. Remote-first: Join a distributed team across LatAm and the U.S.