> Markdown version of [/jobs/ext/3135643-applied-data-scientist](https://www.wearedevelopers.com/jobs/ext/3135643-applied-data-scientist). Every page supports `.md` or `Accept: text/markdown`. Links point to the HTML versions so they work for humans too. Agent guide: [/agents.md](https://www.wearedevelopers.com/agents.md).

---

# Applied Data Scientist

- **Company:** Short, Gram
- **Location:** United States (Remote available)
- **Experience:** Expert
- **Contract:** Permanent contract
- **Skills:** A/B Testing, Data Analysis, Artificial Neural Networks, Data Intelligence, Machine Learning, Marketing Information Systems, Raw Data, SQL Databases, Data Streaming, Feature Engineering, Deep Learning, Information Technology, Machine Learning Operations
- **Published:** September 29, 2026
- **Apply:** https://arc.dev/remote-jobs/j/redirect/pntvqnmem5

## About the Role

* BSc in a quantitative field such as Data Science, Computer Science, Statistics, or Mathematics (must). MSc or PhD is a strong plus.
 * 5+ years of working experience in Data Science or Machine Learning, shipping models to production.
 * Direct, hands-on experience building recommendation and/or personalization systems in production (a firm requirement, not a nice-to-have).
 * Strong applied ML / data science foundations: modeling, evaluation, feature engineering.
 * Neural networks and deep learning (must): hands-on experience training and adapting deep learning models for real-world production use.
 * Time series data (must): hands-on experience modeling temporal data, including forecasting, trends, seasonality, and noisy real-world signals.
 * Ability to take a model or architecture and adapt it to real, messy data - practical, not purely theoretical.
 * Production and engineering chops: get models serving reliably and keep iterating.
 * Experimentation discipline: A/B testing, causal reasoning, honest metrics.
 * Data fluency: SQL and large data warehouses, comfortable in the raw data.
 * Clear communication and a bias to ship.

## Description

* Recommendation & Personalization: Own recommendation and personalization across the entire user journey: which series or episode to watch next, how series are ordered on screen, when and how notifications are sent, when a popup appears mid-viewing, which pricing offer a specific user sees - and more, as new surfaces come online.
 * Predictive Modeling: Turn raw behavioral data into predictive models for retention, churn, and conversion propensity. Predicted LTV (pLTV) is especially critical here - both to power personalization and as a signal we feed back to ad networks to help them find higher-quality users.
 * Applied Model Adaptation: Take existing models and architectures and adapt or fine-tune them to our data - applied, not from-scratch research.
 * Production Ownership: Ship models to production and own them end-to-end: serving, monitoring, retraining, and iteration.
 * Experimentation: Design and read experiments (A/B, causal) to prove real business lift, not offline metrics alone.
 * Cross-Functional Partnership: Partner with Product, Content, and Growth to turn model outputs into decisions.
 * Beyond Rule-Based Analytics: Replace hand-tuned heuristics with models that learn., * Recsys at scale: retrieval / ranking, sequence or session models, embeddings.
 * LTV, churn, or propensity modeling in a consumer / subscription product.
 * Streaming, entertainment, consumer / mobile, or growth-marketing data.
 * MLOps: feature stores, model serving, monitoring, retraining pipelines.
 * Causal inference / uplift modeling.