Research infrastructure
Research Engineer
This role lives at the intersection of research and infrastructure. You will build and maintain the systems that make AI visibility research possible at scale — prompt runners, answer capture pipelines, normalization layers, and the internal tooling that keeps quality high as models and assistants change.
Resumo da vaga
Build the systems that run strategic prompts at scale, capture raw model output reliably, and feed clean data into evaluation pipelines.
Por que esta vaga existe
The research pipeline is the foundation everything else sits on. Without reliable prompt execution and output capture, the product has nothing to evaluate. We need someone who treats the pipeline as a product, not just plumbing.
Primeiros 90 dias
Map and document the current prompt execution and capture pipeline end to end.
Por que esta vaga existe
The research pipeline is the foundation everything else sits on. Without reliable prompt execution and output capture, the product has nothing to evaluate. We need someone who treats the pipeline as a product, not just plumbing.
No que você vai trabalhar
- Design and operate prompt execution systems across multiple AI assistants at scale.
- Build output capture, normalization, and storage layers that handle variability in model behavior.
- Own the reliability and observability of the research data pipeline end to end.
- Work closely with the research team to translate evaluation requirements into efficient system behavior.
Como é um bom encaixe
- Strong systems and backend engineering instincts applied to data pipelines and async workloads.
- Experience building reliable, observable systems that handle high cardinality and partial failures gracefully.
- Comfort reasoning about LLM output formats, rate limits, latency, and prompt execution logistics.
- A product mindset: you treat the pipeline as something users depend on, not just infrastructure.
O que vai te empolgar aqui
- Building the infrastructure that makes serious AI research operationally possible.
- Owning reliability for a data layer that directly determines product quality.
- Working on a system that touches every corner of the product.
Primeiros 90 dias
- 01Map and document the current prompt execution and capture pipeline end to end.
- 02Identify and fix the top two reliability or observability gaps in the research data path.
- 03Ship at least one improvement that materially reduces the cost or latency of a core research workflow.
Processo seletivo
O processo é intencionalmente curto, direto e ancorado no trabalho real.
- 1
Candidatura
Envie seu histórico, trabalhos relevantes e por que essa vaga faz sentido para você.
- 2
Conversa inicial
Uma conversa focada no seu trabalho, no seu julgamento e na vaga.
- 3
Aprofundamento específico da função
Uma conversa ou exercício que se parece mais com o trabalho real do que com um loop genérico.
- 4
Conversa com fundador
Uma conversa final sobre padrão, ambição e como o sucesso se pareceria aqui.
- 5
Decisão
Fechamos o loop com clareza e andamos rápido quando existe convicção.
Precisa de contexto antes de se candidatar? [email protected]
Research Engineer
A vaga já está visível no site. As candidaturas liberam assim que a vaga correspondente no Dover estiver ativa.
As candidaturas ficam fechadas até que a vaga correspondente no Dover seja ativada. Enquanto isso, você pode escrever para [email protected].