Dear Spotify Data Science Team,
I am applying for the Data Scientist — Personalisation role at Spotify. Music recommendation is one of the most challenging and fascinating applied ML problems, and Spotify's investment in doing it well is exactly the environment I want to work in.
At Meta, I worked on the Feed Ranking team, designing and running causal experiments to measure the downstream impact of ranking changes on long-term user wellbeing — not just short-term engagement. I built a difference-in-differences framework to separate organic behaviour from nudged behaviour, which became the team's standard methodology. I also fine-tuned a transformer-based content classifier that improved harmful content recall by 31% with no precision degradation.
My PhD research in NLP (MIT, 2021) focused on cross-lingual transfer learning, giving me a strong foundation in both the theory and practice of modern ML. I'm excited by the prospect of applying sequence modelling techniques to audio and listening behaviour — a domain where the signal richness is extraordinary.
Thank you for your consideration. I'd love to discuss how my background in causal inference and recommender systems could contribute to Spotify's personalisation mission.
Sincerely,
Arjun Mehta