Data Scientist (Special Projects)

Do you check the maths before launch like Katherine Johnson in Hidden Figures, listen to the noise until you find the signal like Ellie Arroway in Contact, and, when a record is missing from the archive, go and find out why like Obi-Wan in Star Wars?

Twinkle helps people find films and cultural experiences worth their time. A robust film dataset lies underneath to make that possible. With a small team of freelancers, we are enriching that dataset across metadata dimensions to get accurate, trustworthy data that helps guide our users to the most suitable film for their taste, mood and setting.

Join us at the forefront of entertainment innovation! We at Malao are on a mission to transform the entertainment industry, creating inspiring stories with positive role models that spark change in our world. Across innovative technology platforms, original films and unique location-based experiences, at Malao you’ll dive into a game-changing environment where we combine human intuition with artificial intelligence to bring the world more Meaningful Entertainment.

Twinkle is our B2C entertainment discovery app that aims to become the go-to destination for users overwhelmed by content choices: a cinematic compass that combines the best of artificial intelligence with human intuition to transform how people find and share films they’ll love. Everything it does depends on one foundation: deep, high-quality film data. With support from investors, Creative Europe and the European Institute of Innovation and Technology, EIT Culture & Creativity, we are now completing that foundation and bringing Twinkle to the wider market.

• Building and expanding the film metadata layer that powers Twinkle — taking our current dataset from around fifteen fields per film to fifty or more nuanced, structured fields that unlock intelligent discovery, filtering, and AI-driven recommendations
• Finding the right source for each field, from public datasets such as TMDB and Wikidata to partner data and primary sources, and making a sound call on which to trust when more than one is available
• Running and improving the AI-assisted pipelines that generate these fields at scale, including the prompting strategies, validation logic and edge-case handling that keep quality consistent for well-known and obscure films alike
• Cleaning, normalising and structuring what comes back so it can power production systems reliably
• Working with our full stack engineering team so the enriched data lands in Twinkle in a form that can be queried and surfaced to users

• A competitive rate for the duration of the project or a short-term employment contract if you prefer
• Access to any AI and productivity tools you need to do your best work
• A great office in the heart of Munich’s innovative business quarter, with the flexibility to work remotely
• ... and being part of a project team building products with the potential to make a real, positive difference in our world

• You are available full-time between 16 November and 18 December, as a freelancer or on a short-term contract
• You have hands-on experience working with large, messy, real-world datasets, knowing how to source, clean, structure, and validate data at scale, and you don’t need perfect inputs to produce reliable outputs
• You are fluent in Python and comfortable with the data science stack, i.e. pandas, NumPy, and similar tools are second nature, and you know how to move from exploratory analysis to something production-ready
• You have practical experience using LLMs and prompt engineering to generate structured data, you think carefully about prompting strategy, understand the failure modes, and know how to design for consistency and quality at scale
• You are curious about film and entertainment data, you understand why metadata quality matters and you are excited by the challenge of capturing something as nuanced as narrative, tone, and genre in structured form
• Malao’s values of truth, justice, courage, gratitude and joy resonate with you, and you thrive in an environment of open communication and candour