About the Role
Gratia Health is looking for a self-motivated mid-level Product Analytics Engineer to own our data landscape. As the sole data practitioner, you will embed directly within our product and engineering team. Because our ingestion pipelines are already automated and lightweight, you won't waste time on infrastructure plumbing. Instead, you will build user behavioral models, design personalization algorithms, and maintain our dbt/Snowflake analytics stack to directly power in-app targeting and content recommendations. You'll work with the entire team to interpret your findings into real product features and vision.
About Gratia Health
Gratia is a venture-backed health technology startup that specializes in helping healthcare employers create healthy, sustainable workforces through its behavior change platform. Our goal is to develop and deploy a portfolio of data informed programs that help create more sustainable work environments for shift-based healthcare providers, like nurses. Our programs are designed using proven behavioral science concepts to address chronic workforce challenges within systems and are deployed using our proprietary software platform that gamifies the experience for participating providers. The end result is increased compensation and satisfaction for providers and a definitive, significant near-term ROI for employers.
About our Team
You'll be joining the engineering team, currently 3 members. We're all remote and live across the US. We love to collaborate via paring, async tools, and timely gifs. We have created an environment of kindness and generosity. We lift each other up, don't blame, and solve problems when we see them. Every member operates with high levels of autonomy to produce the best work in the way that works best for them and the team. As the only data engineer, you'll have to be self-guided and self-motivated, but you'll have a strong group of very smart engineers at your back. We believe in the Agile Manifesto and do our best to let it guide us. No daily standups, no story points, just working software. Our general disposition is towards doing, not talking about doing. We all work directly with the business, our end users, our clients, and anyone we need to, to build the best product possible.
What You'll Do
- User Personalization & Targeting: Build, evaluate, and refine predictive models—expanding our current relevance scoring model into automated content targeting and user personalization.
- Behavioral Modeling: Apply statistical analysis and machine learning techniques to product event telemetry (PostHog) to understand user engagement patterns across core app mechanics like Compliment Circle, Care IQ, and Shift Pickup.
- Data Transformation & Syncs: Maintain dbt transformation models in Snowflake and oversee the automated sync of model outputs (e.g., relevance scores) back into our production Postgres database via Airbyte.
- Telemetry Design: Partner with engineering to establish clean PostHog tracking schema that serves as training data for ML models and feature analytics.
What We're Looking For
- 2+ Years Experience: Background as an Analytics Engineer, Product Analyst, or Product Data Engineer.
- Modeling & ML Proficiency: Hands-on experience building user-level or predictive models using Python (
pandas, scikit-learn, and exposure to PyTorch or TensorFlow). - Statistical Rigor: Strong foundation in statistics, user segmentation, hypothesis testing, and behavioral data modeling.
- Data Engineering Competency: Advanced SQL fluency, experience writing dbt models in Snowflake, and familiarity with sync-back workflows to production databases.
- Product-Driven Execution: Ability to connect mathematical models directly to product outcomes like engagement, retention, and content relevance.
- Autonomy & Ownership: Comfortable serving as the single data owner on an engineering team, taking initiative without needing dedicated data management.
Visionary Mindset
- Targeting & Personalization Vision: Proactively identifies opportunities to turn user interaction data into predictive features and personalized product experiences.
- Pragmatic Modeling: Balances statistical rigor with product execution—building simple, interpretable models first before jumping into heavy neural architectures.
- End-to-End Ownership: Comfortably takes a model from PostHog event logging to dbt transformations, Python feature engineering, and production Postgres syncs.
Collaborative Force:
- Looks at the data and makes it say something
- Partners with product, engineering, and leadership teams to understand data needs
- Designs and implements data schemas and collection patterns that integrate seamlessly with our application
- Collaborates with engineering team to ensure the application captures the right data in the right format
- Advocates for data-driven decision making across the organization
- Excellent remote communication skills, working effectively with a distributed team
Our Tech Stack
- Ingestion & Storage: Airbyte Cloud, PostHog, Snowflake.
- Transformation: dbt, GitHub Actions.
- Serving Layer: Metabase, Claude MCP, Production Postgres.
Interview Process
As part of our interview process, candidates will be asked to share detailed examples of how they've used data to inform product decisions and company goals. We'll also discuss how you've collaborated with application developers to implement effective data collection strategies.
The hiring process for this role looks like this:
Initial Screening (10 min): You'll meet with Brandon, Director of Engineering, (or another member of the engineering team) to determine if you're a real person and qualified for the role. (Unfortunately, we receive a ton of fake or spammy applicants in an effort to minimize the pain of screening them we're trying this short initial screen.
First Real Interview (30 min): You'll meet with Brandon (or another member of the engineering team) to review your fit and learn more about the role.
Technical Interview (90 min): You'll meet with Brandon and another member of the engineering team. We will spend an hour having a technical discussion. With the ultimate goal of determining if your skillset is the right fit for our team. We will dive deep into your experience implementing data strategies in the past. We want to get an idea of what you plan to do at Gratia.
There are no timed questions, trick questions, or high pressure moments. We've found that a group talented engineers talking in detail about technical topics tells us a lot about your skillset and level of competency. You'll also have an opportunity to ask additional questions about the company, culture, and the team.
Final Interview with Leadership (30 min): You'll meet with one of our company leaders to get to know you and see if you have potential to be a great addition to the team.
Gratia Health is committed to providing equal employment opportunities to all qualified individuals. We are an Equal Opportunity Employer and do not discriminate on the basis of race, color, religion, sex, national origin, disability, age, or genetic information.
In accordance with the Americans with Disabilities Act (ADA), Gratia Health will provide reasonable accommodations to qualified individuals with disabilities. If you require a reasonable accommodation to participate in the application or interview process, to perform essential job functions, or to enjoy the benefits and privileges of employment, please contact
[email protected]
Note: Industry research shows that women and those in traditionally underrepresented groups generally don’t apply to jobs unless they check all the boxes for the role. If you feel strongly that you have what it takes for this role but don’t check 100% of the boxes—that’s okay—we encourage you to apply anyway and highlight what you can bring to the table.
And finally since you made it all the way to the bottom, I (Brandon) wanted to let you know we aren't using ai screening tools for this job posting. I'll be reading every single application the old fashioned way. Please write your responses in your own words too. The hardest part of hiring these days is wading through the ai slop, your best bet is to write in your own voice so we can connect as humans.