Kepler Labs→
Data Scientist - Product & Visualization at Kepler Labs · San…
Skills
Job Description
Why this role
We're building the world model for a resilient global economy.
The global economy runs on physical assets: fabs, plants, mines, warehouses, data centers, pipelines. A drought, a port closure or a failed supplier moves through those assets into company earnings, and from there into markets. Our mission is to make the global economy more resilient by translating the physical world into financial impact.
We start with asset-level intelligence for financial markets. Kepler maps the physical assets of every company we cover, models what each one produces and earns, and connects them to the company's financials and to each other — who supplies whom, where revenue concentrates, and what can be substituted.
On that foundation, we layer risk. Physical risk is the first layer, and it's live today: drought, heat, flooding and wildfire, turned into asset-level, dollar-denominated, point-in-time Earnings-at-Risk built to sit next to a Bloomberg feed on an investor's desk. Supply-chain dependencies come next, then geopolitical concentration, with more layers to follow.
We're already working with 3 of the world's top asset managers (over $30 trillion AUM).
Our data science team builds the models: the hazard modules, the revenue attribution, the backtests that show the signal is real. What comes out is a dataset nobody else has. Then it has to reach a person: a portfolio manager scanning the platform, a quant reading a methodology page, a risk team opening a case study, an investor reading our El Niño analysis on their phone. Today that last step is done by hand, by the founders and the researchers, in between everything else. The platform has one engineer. The charts come out of notebooks. The published research is designed on the way out the door.
This role is the data scientist who owns that last step. You'll do the analysis that turns asset-level Earnings-at-Risk into the views a portfolio manager uses, the figures a research page is built on, and the case study a customer forwards to their boss. You'll work with the same data and the same rigor as the rest of the data science team, and you'll build the interface yourself. Half analysis, half product, all of it in front of the most demanding audience in the world.
What you'll own
The analysis behind the product. How asset-level Earnings-at-Risk rolls up to a company, a sector, a portfolio; what a screener ranks on; what a watchlist should flag when a new forecast drops. You decide what a portfolio manager sees, then build it with our CTO: watchlists, screeners, sector and hazard drill-downs, designed from real usage data and onboarding feedback, for people who will never write a Snowflake query.
Published research. Our monthly research pages and white papers. When a new forecast lands you rerun the analysis, update the exposure maps, the charts and the "what to watch" lists, and set the standard the figures are held to. You own the interactive pieces and the chart library behind them.
Customer analysis. Case studies, demo notebooks and the analysis companions that go out with trial data packs, turned around in days. A figure library so no deck starts from zero.
The consumer's side of the data. The example queries, notebooks and data dictionary that make our Snowflake share easy to use, and the usage read-out that tells us what people actually look at and feeds the roadmap.
Exactly how each of these looks is still being written. You'll have a real hand in writing it.
What we're looking for
You're a data scientist first. Python, SQL and the statistics to back a claim: hypotheses, controls, and the sense to know when a result is too clean. You've done analysis where the answer mattered to someone.
And you build the interface yourself. React/TypeScript with d3, Observable Plot, Mapbox or deck.gl. You've shipped interactive data views that people used, not just notebooks.
Visual judgment. You know which chart answers the question and which one just fills space, and you can work inside a design system.
You write for finance readers: headline first, number second, caveat third.
Curiosity about how portfolio managers, quants and risk teams actually work.
3–6 years in data science or analytics with real product or visualization work, or in data journalism, graphics or quant-desk visualization with real analytical depth.
And you're someone we'll want to build alongside for thousands of hours.
Nice to have (genuinely optional — you don't need all, or any)
- Published data stories or interactive graphics — a newsroom, a research house, Observable.
- Geospatial work: maps, rasters, vector tiles.
- Exposure to financial markets or physical-risk data.
- Product management or design experience; comfortable in Figma.
- Full-stack comfort (Postgres, Snowflake, Vercel).
- We'd rather hire someone very smart and very curious who can grow into the domain than someone who ticks every box.
What you get
- $150K–$180K + 0.3–0.6% equity.
- Health-insurance reimbursement (QSEHRA) and a real commitment to grow benefits at every raise and milestone.
- An awesome office space to work from in San Francisco.
- Ownership of the product side of Kepler's data: what the largest institutions in the world see, from the first screen to the published page.
Our values
Curiosity. Pragmatism. Empathy. Transparency. Care.
Process
- Intro call with CEO · 15 min — mutual fit
- Intro call with CTO · 30 min — your technical background
- Deep dive with the founders · 60 min — your background, role by role
- Working Session + Success Plan · 120 min (2 parts) — take a slice of our data, do the analysis, and build one interactive view and one published figure that answer a question an asset manager would ask; present both, then discuss how you'd succeed in the role
- References
- 1-Day Working Interview — assessing long-term fit together
- Offer
How to apply
Fill in the form here and upload your CV. Questions? Email careers@keplerdatalabs.com