Founding Data Science Lead

Employment Type: Full-Time Location: Remote / Miami preferred

About Us

Atomic is a venture studio that co-founds companies by pairing founders with the best ideas, teams, and resources, and funding those with the most potential. When entrepreneurs co-found with Atomic, they team up with an experienced group of operators who have started dozens of companies and created billions of dollars in enterprise value. Industry disruptors like Bungalow, Found, Hims and Hers, Homebound, OpenStore, and Replicant all started at Atomic along with dozens more. Atomic was founded in 2012 by serial entrepreneur Jack Abraham and has offices in NYC, Miami, and San Francisco with a distributed team across North America.

Position Overview

We’re a stealth-mode, AI-native startup reimagining how e-commerce brands connect with customers across the entire commerce lifecycle, starting with revenue recovery. Our first product is a white-glove recovery tool that transforms drop-off into conversions, unlocking precision, personalization, and zero-party data that informs everything from marketing to product development.

Backed by top-tier investors and operating with significant seed funding, we’re hiring our first wave of builders. We’re now looking for a Founding Data Science Lead to join our team and spearhead the design, experimentation, and optimization of our LLM-driven platform. You'll work closely with engineering and product leadership to shape the agentic workflows at the heart of our customer experience.

This is a rare opportunity for a hands-on data science leader to own the modeling strategy and experimentation culture of a company from day one.

The Role

As our Data Science Lead, you’ll partner with the CTO and core team to define and scale our approach to data, experimentation, and model-driven development. You’ll lead the development of performance evaluation frameworks, inform data pipelines, and drive insights from AI trace behavior and customer usage patterns to help us build smarter, more effective systems.

Responsibilities

• Own experimentation strategy for LLM-powered agents, designing and deploying pipelines for A/B testing, evals, and real-time feedback loops.
• Build and refine frameworks for analyzing AI trace behavior, surfacing model weaknesses, and iterating on improvements.
• Define key performance indicators (KPIs) that bridge product goals with AI behavior, including quality, latency, coverage, and conversion metrics.
• Collaborate with engineers and product leaders to shape how data is collected, structured, and leveraged across our systems.
• Make principled decisions around data context and grounding, helping optimize what information is passed to LLMs and why.
• Partner with domain experts and UX teams to translate customer feedback and human workflows into measurable improvements.
• Build dashboards and reporting tools that help the team monitor performance and prioritize areas of investment.
• Stay abreast of emerging LLM trends and AI evaluation research, and prototype new methodologies that can enhance product reliability and accuracy.
• Drive a culture of fast, rigorous experimentation while balancing scrappiness with statistical soundness.

Who You Are

• A hands-on data science leader who loves solving open-ended problems with rigorous experimentation and strong intuition.
• You’ve worked with LLMs in production and have a deep understanding of their quirks, edge cases, and data needs.
• You have a strong grasp of how to build and tune eval systems for agents, prompts, and pipelines where ground truth is often fuzzy.
• You're comfortable with ambiguity, eager to work in fast-paced startup environments, and able to find signal in noisy, high-dimensional data.
• You have experience defining success metrics and building dashboards or analytics infrastructure that supports iterative learning.
• You’re a collaborative partner to engineering and product, with a clear point of view and a humble, pragmatic approach to decision-making.
• You understand trade-offs bet