Noota is the #1 AI platform for meetings in France. We help top companies automate interviews, notes, and hiring workflows using GenAI.
Our next chapter?
We're building a modular agent framework to power complex recruitment and sales workflows.
This is where you come in.
You’ll be the architect and builder of our internal agent system:
Not just writing code, but thinking in systems, behaviors, flows, and prompts
Designing a multi-agent architecture where agents reason, talk, coordinate, and act
Owning the loop from prompting and tool invocation to output parsing and scoring
Think CrewAI, LangGraph, OpenAgents — we’re inspired by them, but we want something tailored to real-world verticals like HR and sales.
Core Missions:
Design and implement the agent orchestration framework
Build custom tools (e.g., ATS connectors, web scrapers, reasoning utilities) the agents can use
Prototype, test and iterate prompts in live environments
Collaborate with product to transform user goals into agent workflows
Make sure the system is scalable, traceable, and production-ready
We’re looking for a builder of agentic systems, halfway between prompt engineer, product-minded dev, and AI architect.
Must-have:
Experience with agent frameworks: CrewAI, LangGraph, AutoGen, etc.
Very solid Python skills, but prefers designing behaviors and reasoning over low-level infra
Proven ability to ship production-grade AI flows
Understands how to create robust prompts, retries, fallbacks, scoring, evaluation logic
Can think in terms of jobs-to-be-done, not just models and APIs
Obsessed with speed, iteration, and impact
Nice-to-have:
Familiar with vector DBs, tool use, retrieval-augmented generation (RAG)
Experience in startups or fast-paced AI teams
You get to own a core part of the AI stack in a fast-growing company
Work directly with the CEO, CTO and Product team, zero bureaucracy
Full access to best LLMs, tools, budget to experiment
Flexible setup: remote, freelance or full-time, your call
