Description:

Quadrivia is a health technology company that has developed Qu, a customizable AI assistant designed for healthcare professionals.
The AI aims to address the shortage of healthcare workers by providing real-time support for clinical tasks.
The role involves owning and evolving the core “brain” service that powers Qu, which includes designing, building, and operating multi-agent LLM systems for real-time communication.
Responsibilities include managing architecture, SLAs, latency budgets, and error modes for Qu’s brain service.
The position requires expertise in low-latency communications, including streaming text and voice, and familiarity with technologies like WebRTC and SIP.
The role also involves multi-agent orchestration, reasoning and optimization techniques, programmatic prompt optimization, and RAG engineering.
Evaluation and observability tasks include validating inputs, verifying retrieval quality, and conducting automated task evaluations.

Requirements:

Candidates must have 5+ years of experience in ML or backend engineering, with a recent focus on LLM systems.
Proficiency in Python is required, along with strong skills in FastAPI, asyncio, pydantic, and production observability.
Experience in building or integrating low-latency text/voice systems is essential, with familiarity in technologies like LiveKit or Pipecat.
A working knowledge of agent patterns and evaluation-driven development is necessary.
Hands-on experience with ReAct and Chain-of-Thought methodologies is required, along with a pragmatic approach to Tree-of-Thought/Graph-of-Thought tradeoffs.
Prior experience in a startup environment is preferred.

Employees will work on cutting-edge real-time agent technology with a top-tier team in the health tech sector.
The company offers enjoyable off-site events in Barcelona.
A high-tech laptop and ergonomic development setup are provided.
There is flexibility in work arrangements, allowing employees to work from home or in a hybrid model in Barcelona or London.

Skills