> Markdown version of [/jobs/ext/4635573-machine-learning-engineer-applied-research](https://www.wearedevelopers.com/jobs/ext/4635573-machine-learning-engineer-applied-research). 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).

---

# Machine Learning Engineer, Applied Research

- **Company:** Whatnot Inc.
- **Location:** New York, NY, United States (Remote available)
- **Experience:** Expert
- **Salary:** $207,000 - $290,000
- **Contract:** Permanent contract
- **Skills:** Python (Programming Language), Machine Learning, Open Source Technology, Recommender Systems, Tensorflow, SQL Databases, Reinforcement Learning, Pytorch, Xgboost, Machine Learning Operations
- **Published:** November 8, 2026
- **Apply:** https://www.juju.com/job/28_25303_175e88e83c14dba0d79ea940766a208a

## About the Role

* 5+ years of industry experience building and deploying ML models to solve user problems at scale.
 * Deep expertise in at least one area: recommendation systems, causal inference, off-policy evaluation, reinforcement learning and bandits, auction or mechanism design, or marketplace experimentation.
 * A track record of applying scientific methods to real-world problems using consumer-scale data.
 * Advanced proficiency in Python, SQL, and common ML frameworks such as PyTorch and XGBoost.
 * Strong grounding in applied statistics, experiment design, and theoretical machine learning.
 * Strong communication and leadership skills, including the ability to influence roadmaps and align cross-functional teams in a remote environment.
 * Preferred: experience in two-sided marketplaces, ads and auction systems, or pricing.
 * Preferred: experience building simulators or economic models of platform behavior.

## Description

* Lead applied research projects focused on marketplace dynamics, including simulation, auction and allocation mechanics, long-term objective modeling, exploration and information value, and marketplace experimentation methods.
 * Drive work from hypothesis to production through literature review, prototyping, offline validation, shadow testing, and shipping online experiments through partner teams in Discovery and Seller.
 * Build system-level models of how the marketplace behaves, such as learned simulators that estimate segment-level impacts from ranking and policy changes, plus surrogate models for long-term marketplace outcomes.
 * Model real market mechanics, including auction and bidding dynamics and discovery exposure allocation as a portfolio problem, with allocation to rising sellers.
 * Improve how a multi-sided live marketplace evaluates changes using off-policy evaluation, switchback and interference-robust experiment designs, and variance reduction.
 * Help strengthen Whatnot's external presence through publications, open-source work, and public benchmarks.