Forward Deployed Engineer

TLDR

• Role: full-stack, embedded on-site deployment of Parsio at industrial clients, reports to CTO

• Key work: ERP/PLM/legacy data integration, production code, turning client edge cases into reusable platform primitives

• Profile: pragmatic full-stack + strong business acumen + client-facing, high autonomy under ambiguity

• Constraints: 25-50% client travel, operational French + English

• Package: €80-90k + equity, Paris hybrid

Who we are

European manufacturing is under pressure. Margins are thinner, supply chains more volatile, competition more global. Yet the largest cost lever most companies have — 70–80% of product cost is locked in engineering decisions stays invisible to procurement teams working from spreadsheets and tribal knowledge.

Industrial procurement is worth trillions of euros. It's one of the last great domains where business-critical data still lives in PDFs, scanned drawings, and heterogeneous spreadsheets. Mid-market industrial manufacturers are world leaders in their niches and the backbone of European competitiveness. But the procurement software they've been handed was not designed for the engineering complexity of industry, and it doesn't serve them. They run real manufacturing complexity on tools designed to buy laptops and pens.

Parsio exists to change that. Legacy tools manage transactions. Parsio optimizes what you buy.

Why now

What we're building wasn't possible six months ago. The bottleneck was never insight but capacity. That constraint just disappeared. This is an 18-month window to build the category winner in AI-native procurement intelligence for industry. The companies and the team that move fastest now will lock in structural advantages the incumbents can't catch.

The role

AI is easy to try. Making it work inside an industrial enterprise is a different problem entirely.

That's the job. You take clients beyond the POC and into production, building the last mile.

You're a full-stack engineer who works directly inside the customer's environment alongside their buyers, engineers, and operators building the production systems that turn Parsio from a promising demo into the tool their sourcing decisions actually run on.

You operate like a startup CTO embedded in the account: you own end-to-end execution of a high-stakes deployment. In practice, this means you:

• Write real production code

• Integrate with legacy systems and brittle data pipelines: ERPs, PLMs, supplier portals, twenty years of Excel.

• Model messy institutional workflows: map how this specific client actually sources, costs, and decides.

• Build specialized skills, workflows, and integrations on top of Parsio: making our agents behave like a highly trained teammate for this client, not a vanilla chatbot.

• Sit alongside analysts, operators, and decision-makers and adapt continuously as requirements evolve.

This work is intentionally uncomfortable. It exposes you to edge cases, organizational friction, and failure modes that never surface in a controlled product environment.

The feedback loop

You are the closest person to the customer, and what you learn flows straight back into the platform:

1. Deploy into a real institution.

2. Solve a concrete operational problem that delivers measurable savings.

3. Identify the recurring structural challenges across clients.

4. Abstract those challenges into reusable primitives.

5. Incorporate them into the core Parsio platform.

The goal isn't to ship one heroic deployment. It's to make the next deployment cheaper because of what you fed back from this one.

The stack

Data stack

• LLM & OCR: Pydantic AI, Logfire, GCP Vertex (Gemini, Claude Sonnet & Haiku)

• CAD analysis: FreeCAD

• ELT: dlt, dbt + ML with Bruin and Scikit-learn

• Orchestration: Python scripts run locally

• Hosting: GCP, GCS

• Database: PostgreSQL (Neon)

Application

• React SPA

• Application backend: Hono (REST API, Drizzle)

• Second LLM backend (separation of concerns): FastAPI + PydanticAI

• Hosting: GCP Cloud Run

Who you are

We hire curious hustlers. This work it requires high agency, insatiable curiosity, and a healthy disregard for the status quo.

Must-have

• Full-stack and pragmatic. You can move from a React fix to a SQL pipeline to an API integration

• Business acumen. You walk into a manufacturer, sit with the buyers and the engineers, and identify where the real money and the real friction and then connect that diagnosis to a product hypothesis

• Customer instinct & client-facing by default. You build trust quickly at every level of an organization communicate with clarity, and know how to tailor your depth to the room. This is a deeply on-site role: you'll spend a large share of your time at clients.

• Owner mindset under ambiguity. You take a vague client problem and drive it to a production system that delivers value.

• Willingness to travel. This role requires significant on-site presence with clients (expect 25–50% travel).

• Operational French and English (international clients).

Bonus

• Real industrial exposure (manufacturing, procurement, automotive, aerospace, steel…).

• Familiarity with CAD / STEP files (cadquery, FreeCAD).

• Experience integrating with ERP / PLM systems.

• You've shipped a typed product taxonomy or ontology (eCl@ss, UNSPSC, or an internal one).

• Experience with LLM evals.

How we work

• We start with a small, synchronous, trust-first team. No org chart between you and the founders; daily standup, decisions made out loud. On your scope, we expect autonomy and initiative.

• Be there in person. Pilots are real clients, not design partners in name only. You'll go on-site to drive adoption, read the internal power dynamics, and watch them actually use Parsio.

• AI-native by default. Claude Code is part of how we ship. It's what lets a team our size be faster and more agile than the big incumbents, and capture the lion's share in a market where speed makes the difference.

Expected behaviors

• Owner mindset. This is your deployment. Think and act like it.

• Bias for action. Ship early, iterate fast, learn from the client.

• Pragmatism. Elegant where it matters, scrappy where it counts.

• Candor. Speak openly, give and receive feedback with humility.

• High standards. Never compromise on quality that impacts customer trust.

What success looks like

The job has two outputs: a client getting real value, and a product that's better for having deployed there. The win is turning the client’s hardest problems into primitives the whole platform inherits is. That's why we also measure as abstraction extracted.

• By month 1: You're embedded on-site with your first client, you understand how they actually source and decide, and your first integrations are live in production.

• By month 3: Parsio runs reliably end-to-end inside that client from their raw files and systems to typed data driving real sourcing decisions with the monitoring, tests, and auditability you put in place.

• By month 6: Your deployment delivers savings the client can point to, and you've turned your first recurring patterns into reusable primitives that ship back to the core platform.

• By month 12: You run multiple deployments, the playbook and primitives you built make each new one faster and cheaper, and you're trusted to make the architectural calls on your scope.

Career path

Forward Deployed Engineer → Senior FDE → Lead / Staff FDE

You join owning a deployment end-to-end. As you grow with the company, you take broader ownership across accounts, turn field learnings into platform standards, set the engineering standards, and mentor the FDEs we hire next.

What's in it for you

• Own a piece of the upside. Founding-level equity in a category-defining company.

• Build something unprecedented. Operationalize the first AI-native sourcing intelligence engine for industry.

• Massive scope & impact. You don't just deploy the product; what you learn in the field defines it.

• Direct customer impact. See your code deliver millions in measurable value.

• Learning velocity. Work at the frontier of AI + industrial data systems, inside real factories.