[Hiring] Junior AI Full-Stack Developer @Concurrences

Role Description

You own RAG infrastructure for 330–430K legal documents. Responsibilities include:

• Chunking pipelines
• Vector database tuning
• API components
• Concrete work includes:
• Intelligent chunking for French court decisions (preserve precedent hierarchy)
• Weaviate integration, embedding strategy optimization, semantic search at scale
• RAG components: retrieval → reranking → context → LLM calls
• Evaluation infrastructure so domain experts measure correctness, not just plausibility
• Ship 3–5 features in month one

Tech stack: Python, FastAPI, Weaviate, PostgreSQL, GCP. Plus: CrewAI/LangGraph/LlamaIndex, Pulumi/Terraform.

A Typical Week

• Monday: Debug chunking pipeline mangling PDFs. Propose fix, pair, ship it.
• Tuesday: Embedding model struggles with French legal terms. Document, propose mitigation.
• Wednesday: Design caching strategy for API. Watch latency drop.
• Thursday: Code review. Close issues. Plan sprint.

Who You'll Work With

• Anton Bossenbroek (CTO): 20+ years at BCG, Bain, EY, Deloitte, and C3.ai. Your primary technical collaborator on architecture and AI integration.
• Nicolas Charbit (founder): 20 years building Concurrences into the definitive antitrust platform. Shapes domain direction and product strategy.
• Small, focused, high-ownership team. This is a strategic AI initiative — you'll matter immediately.

Who This Role Fits

• You ship real systems.
• You reason from first principles.
• You're a generalist ready to learn fast — APIs, databases, infrastructure, legal domain, all at once.
• Motivated by collaboration, mission, and growth.
• You ask until you understand.

Requirements

• Python: Non-trivial production code. Async, testing, error handling.
• Backend/API: Built and deployed a web service (FastAPI, Flask, Django).
• Linux & Git: Terminal-comfortable. Workflows, shell scripts, SSH.
• SQL: Write queries. Understand joins and indexes.
• Intellectual honesty: Admit unknowns. Growth mindset non-negotiable.

Big Pluses

• Vector databases or RAG (Weaviate, pgvector, Pinecone)
• Agentic frameworks (CrewAI, LangGraph, LlamaIndex)
• GCP or cloud platforms
• Infrastructure as Code (Terraform, Pulumi)
• Law, legal tech, court decisions
• French language (A2+ valuable)

Compensation & Timeline

• R$6,600/month (~€1,200), paid EUR via Wise.
• Independent contractor, monthly invoicing.
• Initial 3-month project engagement with clear path to ongoing collaboration at a higher rate.
• Start: ASAP, within 4 weeks.

Interview Process

• Application — CV, GitHub, three written answers (submitted via application form below).
• 15-min intro call with Anton (CTO) — Who you are, what excites you, fit.
• Two 30-min deep dives — Architecture & trade-offs. Real RAG problem-solving.
• 20-min conversation with Nicolas (founder) — Domain understanding, alignment, culture.
• Timeline: 2–3 weeks, application to offer.

Remote & Async

• English required. French a plus.
• Remote-first, async work culture.
• The CTO is based in Vancouver, the company in Paris.
• We coordinate across time zones — responsiveness matters more than fixed hours.