Data & Analytics Lead

Function: Data & Analytics
Reports to: Founder & CEO, working alongside our CTO and President of Technology
Location: Los Angeles — hybrid, monthly in office. Southern California strongly preferred; exceptional fully remote candidates considered - see Location section for details.
Employment type: Full-time, exempt
Base salary: $120,000 – $180,000, depending on experience and location — see Compensation for details
About Cabrella
Cabrella insures goods in transit — from general commodities and everyday merchandise through to the shipments most carriers would rather not touch: loose diamonds and finished jewelry, precious metals, graded trading cards, fine art. Mainly parcel, and increasingly freight and other conveyances. We do it with software we build ourselves: a native TMS with coverage inside the shipping workflow, an insurance API partners embed in their own flows, and a white-labeled claims platform.
About the Role
We are at the beginning of this. Today we have Azure, an ETL process, and a handful of Power BI reports. That is genuinely it. Everything described below still needs to be built, and we are hiring the person who is going to build it.
This is a hands-on job. You will be the most senior data person at Cabrella, reporting to the founder — but that means you write the SQL, build the model, and fix the pipeline, not that you delegate it and review slides. If you are looking for a role where someone else touches the data, this is not it.
You will not be alone. We have offshore capacity that produces good work when it is well specified, a DBA, and a technology team you will work alongside. Part of the job is directing that capacity well — writing clear specs, reviewing what comes back, and sending it back when it is wrong. But the architecture, the models, and the standards are yours to build.
The first job is a foundation. One place where our own operating data — the TMS, policy, claims, quoting — lives alongside the outside data that gives it meaning: carrier scan and exception feeds, partner API traffic, carrier performance. Those live apart today, which is why nobody can answer a question that crosses them.
The second is making the numbers trustworthy. One definition of every metric the company currently argues about, and reporting that leadership, underwriting, claims and account management rely on without checking it twice.
The third is the one that pays for the role. We insure parcels that carry real value, and loss trends move fast. If a particular carrier, lane or warehouse starts going bad, catching it on day 10 instead of day 50 can be the difference between a profitable book and one that loses money. Building the alerting that catches it does not require machine learning. It requires data you can trust and thresholds on top of it. That is squarely this role.
What You Will Own
The data foundation. A warehouse or lake that brings our internal systems together with external sources, modeled so questions that cross them can be answered. Ingestion, transformation, lineage, and a sequenced plan to get there.
Metrics everyone agrees on. One definition per measure, documented, certified, and defended when someone wants their own version.
Reporting people trust. A governed semantic layer and self-serve model that reduces ad-hoc requests rather than multiplying dashboards. We are on Azure and Power BI today; we are open on tooling if you make a good case.
Data where people already work. Our teams should not have to ask for a report, and they should not have to leave our own platform to get an answer. A large part of this job is preparing and exposing data so it can be surfaced inside our admin and claims systems — whether that is embedded reporting or our application querying a governed model through an API, with row-level security so each user sees only what they should. You will make that call with our CTO and President of Technology and deliver it with their team.
Anomaly and loss-trend detection. Alerting on the metrics that matter — loss frequency and severity by carrier, lane, commodity and partner — routed to someone who can act on it.
Offshore delivery. Scoping and directing analysts and engineers. Written specs, review cadence, quality gates, and accountability for what ships.
The security boundary. Access control, environment separation, masking where it is needed, and audit trails. We handle PII, payment data and high-value declared amounts, and we hold SOC 2 — keeping it is part of this job. Contractor access is designed and documented, not assumed.
What We Are Looking For
Required:
You have built a warehouse or data lake with your own hands. Not overseen one. Ingestion from multiple systems, transformation as version-controlled code, orchestration, a modeled layer people query. Name the platform on your resume — Redshift, Snowflake, Databricks, BigQuery, Synapse or similar — and tell us what the sources were and what broke.
You have joined internal operating data to external sources. Vendor feeds, partner APIs, purchased datasets — and dealt with the reality: inconsistent keys, late files, schema changes nobody warned you about, and matching records across systems never designed to agree.
You have personally built a governed metric or semantic layer. Certified datasets, definitions that hold up, a self-serve model, and row-level security enforced at query time. Be ready to describe a metric definition dispute you settled and what you rationalized away.
You have delivered data into an application, not just a BI tool. Embedded reporting, or an API layer a product team consumed — something where the end user got their answer inside the system they were already using. Working alongside engineers to ship it counts; you do not need to have written the front end.
Strong hands-on SQL and working Python. You write both regularly and would be comfortable doing so from week one. This is not a review-and-approve role.
Some experience directing contractors or an offshore team. You have written a spec someone else built from, reviewed the result, and rejected it when it was wrong. Bring a spec you wrote. You do not need to have run a large vendor organisation.
You have implemented access controls on sensitive data. Not just complied with someone else’s. Masking or tokenization, environment separation, least privilege, audit trails. We hold SOC 2 and intend to keep it.
Judgment about sequencing. You can look at a messy estate and say what has to be true before the next thing is worth attempting — and hold that line when someone senior wants the dashboard now.
Strongly Preferred:
Insurance, shipping, logistics, parcel or supply chain. Carrier and tracking data, policy and claims models, loss ratio, frequency and severity, subrogation — and how dirty this data is in the wild.
Amazon Redshift. Hands-on Redshift experience is a significant plus and should be visible on your resume — cluster and workload management, distribution and sort keys, Spectrum, and loading at volume.
Microsoft and Azure. Azure SQL, Power BI, Fabric a plus. Equivalent depth on Snowflake, BigQuery or Databricks transfers.
Fraud or predictive modeling exposure is a plus, not a requirement. We intend to add that capability later, under this role’s direction — likely as a specialist engagement. Someone who has shipped models before can scope it and judge the vendor. Someone who has not can still do this job well.
Deliberately not required
Production machine learning. Fraud models. Conversational AI over data. Enterprise data strategy owned at board level. Those things matter to us and they are not what this job is, because every one of them depends on a foundation that does not exist yet. We would rather hire someone excellent at building that foundation than someone whose best work sits three layers above where we are.
Where it leads. What we eventually want is pre-transit risk scoring — telling a customer, before they ship, how a lane, carrier and commodity combination is likely to perform. We do not do that today in any real way, and we will not get there without the work described above. If that is the kind of problem you want to be pointed at in two years, this is the job that earns the right to it.
Compensation
Base salary $120,000 – $180,000.
That is a wide range on purpose, and here is what moves you within it: where you live, and how much of this you have done before. A Southern California hire who has perso

Work arrangement
Yes