Lead Analytics Engineer - Environmental Intelligence

The fashion industry is responsible for 5-10% of global greenhouse gas (GHG) emissions. As climate pressure grows, fashion companies need carbon tools built for their reality, not generic platforms.

Carbonfact is the Environmental Intelligence Platform for textile and fashion. We turn messy business data into environmental intelligence so brands can measure, model, reduce, and report on their impact with confidence.

We've raised $17M from top-tier investors including Alven, Headline, and Y Combinator, and we're trusted by brands like Carhartt, GANNI, On, Burton, Armedangels, Jack Wolfskin, and many more.

Lead Analytics Engineer at Carbonfact

Our platform is organized around four pillars: Collect customer data through custom connectors, Measure impact from a single process to an entire brand, Reduce footprint through simulation tools and Report with audit-ready, regulation-compliant outputs.

As our first Analytics Engineer, you will own the layer that makes our numbers trustworthy. The footprint a customer sees, the KPIs our team relies on, and the data our AI agents query all flow through one metrics system. Your mission is to make sure data quality issues never reach customers, and build the function to keep it that way as we scale.

We are scaling our teams, AI is accelerating our pace of innovation, and our data stack has grown faster than its ownership:

• Our warehouse transformations run on lea, an open-source framework born at Carbonfact, whose creator has since moved on.

• Engineering has been maintaining a metrics layer that belongs with Data.

You will join the Data team as an individual contributor, reporting to Félix, our Head of Data. This is a senior IC role: you lead through the quality of your judgment and the vision you set for analytics engineering. You will own the metrics system and work daily with Engineering, Data Science, and Product to keep every number consistent, tested, and documented.

Already on your desk

Real items from our wishlist - the kind of work waiting for you:

• One factory, three data sources. Factories share data through our suppliers platform, brands through the platform, and our parsing layer extracts factory information from customer files. You design the model that reconciles all three, so cross-brand factory analysis becomes one clean query.

• Converge the sources of truth. The same metric can currently be computed in up to five places! You fold them into one governed semantic layer for the platform, the team, and our AI agents.

• Guard the per-account view. A methodology change might look immaterial as a whole while capable of moving one customer's footprint by double digits. You turn per-customer checks into a pre-merge gate, with a cost and runtime budget you define.

• Documentation as a build artifact. Our warehouse has 276 SQL models and no column dictionary. Our parsing layer already generates docs from code; you bring the same discipline to the metrics layer, for humans and AI agents.

What you will do

• Own the SQL transformation layer that materializes our metrics in BigQuery: guidelines, contracts, tests, monitoring, and CI gates.

• Build and govern the semantic layer that becomes the single source of truth for the platform, internal KPIs, and AI tools.

• Set the vision and standards for analytics engineering at Carbonfact: the practices, tooling choices, and technical direction the function will grow into as we scale.

• Implement and test the data contracts behind company KPIs, and challenge definitions that would not survive an audit.

• Partner on product bets: define success metrics that survive scrutiny and implement their tracking on governed models.

• Treat agent-facing tooling as part of the data layer: Claude skills, MCP tools, and generated dictionaries.

• Detect bad data early and route failures to the right owner, before a customer ever sees them.

What you won't do

• Be the ad-hoc query desk. This role exists to converge and govern the metrics system, not answer every one-off number question.

• Build dashboards all day. Dashboard-first BI is explicitly not the job.

• Decide alone what metrics mean. Domain experts own the meaning; you own the engine, guidelines, and checks that keep definitions honest.

• Build heavy ETL. Fivetran handles ingestion and GitHub Actions orchestrates transformations - no Airflow, no Spark clusters.

Who you are

Three traits matter more to us than any tool on your CV:

• You thrive on complexity, yet manage to break things down to make them clear. The fashion industry and carbon accounting in general are more complex than it seems. We’re aiming to give clear environmental insights to customer with various degrees of data proficiency.

• You're comfortable in loosely defined contexts. You find solutions, make trade-offs, and move things forward without waiting for a fully-specified brief.

• You have owned something, not just contributed to it, and were accountable when it broke.

• You treat documentation as a build artifact: generated from source, tested and versioned.

And concretely:

• 4+ years of professional experience in analytics engineering or an adjacent data role

• SQL proficiency on a columnar warehouse, BigQuery ideally

• You have run a transformation layer in production, not just contributed models to one

• Comfortable outside dbt: we run our own open-source framework and value tools chosen for needs, not trends (but why not move to dbt?)

• Enough Python to be at ease in the engine and scheduled jobs - you are not SQL-only

• You communicate well in English and can walk non-data colleagues through a data decision

Bonus points:

• Semantic layer work: MetricFlow, Cube, LookML, or the Open Semantic Interchange direction

• BI-as-code tools like Rill

• Data-as-product experience where numbers ship to customers; audited or regulatory reporting is a strong plus

• GitHub Actions as an orchestrator

If you match most of this but not every line, apply anyway. We hire people, not skills.

What we offer

• A transparent, collaborative, high-agency culture rooted in Carbonfact's principles here

• AI-first tooling: generous access to frontier AI models, Claude Code, GitHub Copilot

• MacBook, headset, and all the modern essentials

• 100% coverage of premium health insurance with Alan

• Fitness (Gymlib) and commuter benefits

• Annual learning budget to support professional development

• Team retreats twice a year and occasional onsite visits to brands

• Transparent compensation framework with level-based salary and equity; promotions tied to impact and performance; the salary range for this role is €66K - €88K per year based on leveling

• Compelling equity package with employee-friendly exercise rights

• Paris-based office, with flexible remote policy (full remote negotiable, depending on profile)

Hiring process

• Submit your application online

• Introductory video call with someone from the Data team (30 min)

• Code walkthrough: reading and debugging warehouse SQL together (45 min, can run back-to-back with the next session)

• Governance session: how you would converge our sources of truth (45 min)

• Principles interview with a Carbonfact co-founder

• Final debrief and reference calls

If there is mutual interest to move forward, we’ll extend you an offer to join our team