About Early Media
Early Media is creating and distributing groundbreaking original series. We craft narratives with viral potential while preserving the heart of indie production, telling meaningful stories audiences actually care about. Every creative decision is sharpened by a proprietary testing and marketing engine that collapses the distance between filmmaker and audience into the tightest iteration loop in the industry.
We're hiring a senior data scientist to own applied data science for Twist and our Analytics Engine — modeling what makes content perform, how audiences convert from discovery to premium viewing, and how early signals predict which pilots are worth greenlighting. This is a hands-on modeling and analysis role where you will work alongside our Junior Data Analyst (pairing and mentoring), partner with our senior data science advisor and operate within the analytical roadmap set by the Director of Research.
What you'll own
Applied modeling for the Analytics Engine — content performance scoring, engagement classification, sentiment analysis, and campaign attribution.
The pilot performance prediction loop — forecasts vs. actual outcomes, calibration, and accuracy tracking.
Modeling of Twist user behavior post-launch — acquisition attribution, watch-depth, retention, and conversion from campaign touch to in-app engagement.
Audience segmentation with real analytical depth, not surface-level demographics.
Mentorship of our data analyst, and the team's overall statistical rigor.
Turning findings into recommendations Producers and leadership can actually act on.
Must-haves
Skeptical of metrics that look good but don't actually predict outcomes.
6+ years of hands-on data science work with real, measurable business impact.
Strong Python and SQL, with a solid grounding in statistics, experimental design, and ML fundamentals.
NLP experience — sentiment analysis, text classification, and comfort working with LLM-based pipelines.
Attribution modeling experience, ideally where attribution is genuinely hard (short-form video, organic social, multi-touch).
Able to take an analytical agenda and run with it without daily direction.
Can explain findings clearly to people who aren't data scientists.
Nice to Have
Media, streaming, entertainment, or creator economy experience.
Recommendation systems or content discovery ML background.
Experience with modern data stacks (Snowflake, dbt, lakehouse architectures).
Experience with LLM evaluation, prompt versioning, or agentic workflow patterns.
Skills Required
- At least 6 years of hands-on data science experience with measurable business impact
- Strong Python and SQL skills
- Knowledge of statistics, experimental design, and machine learning fundamentals
- Natural language processing experience, including sentiment analysis and text classification
- Experience with LLM-based pipelines
- Attribution modeling experience, ideally involving short-form video, organic social, or multi-touch attribution
- Ability to work independently from an analytical agenda
- Ability to communicate findings clearly to non-data scientists
- Media, streaming, entertainment, or creator economy experience
- Recommendation systems or content discovery machine learning experience
- Experience with modern data stacks, including Snowflake, dbt, or lakehouse architectures
- Experience with LLM evaluation, prompt versioning, or agentic workflow patterns
What We Do
Early Media is a production and distribution company creating original series designed for viral potential. By combining indie production values with a proprietary testing and marketing engine, they aim to bridge the gap between filmmakers and audiences, bypassing legacy media bureaucracy to deliver meaningful stories through modern, high-impact storytelling platforms.
