_*]:min-w-0 gap-3 !gap-3.5">
Your RoleWe're hiring a Senior Data Scientist to own measurement across the surfaces that move our customers' money. Reporting to the Head of Data, you will settle the metric definitions that three teams currently disagree on, set the thresholds that decide when a human touches a document, and stay accountable for those thresholds after they go live. We cut a standalone data scientist role to fund this one. Building a model got cheap; being answerable when a threshold is wrong did not.
Responsibilities will includeObjective #1: In your first 30 days, you will:- Get deep into Ascend's products, data infrastructure, and the metric definitions the three teams currently disagree on
- Build relationships with Product, Engineering, and the Head of Data to understand how measurement decisions get made today
- Audit the scores and thresholds already in production and map where instrumentation is missing
Objective #2: In your first 60 days, you will:- Propose and settle the accuracy, success, and exception metrics for Ascend's products
- Start owning the thresholds that decide when the system acts on its own versus routing to a person
- Partner with Product and Engineering on instrumentation for at least one upcoming feature before it ships
Objective #3: In your first 90 days, and beyond, you will:- Fully own the accuracy, success, and exception metrics across all products
- Build and ship models that change a workflow rather than describe one: confidence scoring on extracted data, ranking and matching for reconciliation, sequencing for collections and recovery
- Monitor every score in production - predicted against realized, drift, and an explicit re-fit or retire date - and design the experiments our volume can actually power, saying plainly when a surface can't support one
- Turn product usage data into roadmap input
You might be a good fit if you are/have:- Owned product data inside a live workflow rather than reporting on one, and can name the decision your number changed
- Deployed a model into a live workflow and owned it afterward, including calibration, drift, and the call to re-fit or retire
- Comfortable explaining calibration, drift, predicted-versus-realized, and re-fit timing to a non-technical owner
- Strong in SQL and feature engineering
- Experience with payments is a plus
- Someone who thrives in the ambiguity of an early-stage startup and will say when the data is too weak to decide, then propose a next step anyway
- Experience working in our toolset: Slack, Notion, Excel, Front, and Linear
- Strong written and verbal communication skills, ability to quickly understand complex (and sometimes dense) subject matter, and great attention to detail
