Saltar al contenido principal
StrideRecruit

AI Engineer

Koltin

Fuente: RemoteJobs.org — feed autorizado

Ubicación
México
Tipo de empleo
Jornada completa · Remoto
Experiencia
No especificado
Publicado
9 nov 2026
Última vez activo
17 nov 2026
Fuente
RemoteJobs.org
Empleador
Koltin
Ver en RemoteJobs.org y postular ↗

Este enlace abre la oferta en RemoteJobs.org, donde continúa la postulación. No está confirmado que sea el sitio web del propio empleador.

Descripción del puesto

La traducción automática no está activada. La descripción se muestra en su idioma original.

About us

At Koltin, we’re redesigning the way we care for our moms, dads, and grandparents to help them remain healthy and independent for as long as possible.

How do we do it? Through our health memberships, which provide personalized, preventive care to support our members’ long-term well-being.

We are the first company in Mexico offering health memberships that include major medical insurance coverage for older adults-up to 84 years old-backed by BBVA Seguros Salud.

Sobre el rol

We are looking for our first Senior Software Engineer - Agents & LLMs. Our members write to us tens of thousands of times a month, often with fear, anger, or a story behind the question. You will put agents into production to serve them, in a domain where an answer that sounds right and is wrong is the worst possible error.

What you build becomes the foundation for how the whole company builds with LLMs. You work with product, data, and other engineers, and you decide what we build, what we buy, and what we assemble.

Generalist with a strong specialty (T-shaped): you move across the entire stack, with a leaning toward LLMs and agents.

Lo que harás

  • Agents in production. Orchestration, tool calling against our systems, and member context assembled from multiple sources, over WhatsApp and whatever channels come next.
  • The quality bar. Evals on real cases and acceptance criteria defined before launch.
  • The platform. LLM observability (traces, cost, latency, quality), evals as infrastructure, provider-agnostic model access, and prompt versioning, so other teams don't start from scratch.
  • Security in a regulated environment. Guardrails, PII, identity verification, and human-in-the-loop where the risk calls for it.
  • End-to-end projects. You go after the problem behind the request, break it into phases, and take it to production with visibility into risks and decisions.
  • Raising the bar. ADRs, review, pairing, and patterns that other teams reuse.

Lo que buscamos

Contratamos por fundamentos, modelos mentales y patrones. El stack se aprende.

  • Must have a track record of at least 5 years across the full development lifecycle.
  • Have taken LLMs into production, with at least one system facing real users that you operated after launch.
  • Serious tool calling. You have designed tools and know what to do when the model calls the wrong one or with the wrong arguments.
  • Serious evals. Datasets of real cases, LLM-as-judge calibrated against humans, regressions run before every prompt or model change.
  • Datos sensibles en producción: PII, retención, acuerdos de procesamiento con proveedores.
  • Español nativo o fluido, inglés profesional.

Suma puntos

  • Context engineering: compactación, memoria, context provisioning vía tools.
  • Retrieval agéntico y servidores MCP, más allá de RAG sobre embeddings.
  • Agent security: direct and indirect prompt injection, minimal permissions per tool, red-teaming.
  • Multi-agente, subagentes y ejecución durable.
  • Costo y latencia: prompt caching, routing entre modelos, structured outputs.
  • Online evals: shadow mode, A/B tests between prompts or models, and human feedback that flows back into the dataset.
  • WhatsApp, voz, salud o seguros.

Cómo trabajamos

  • We use AI for everything, and you own what reaches production.
  • Self-merge by default, with CI, types, and tests as the gate. Architecture decisions live in ADRs, and that is where we focus review.
  • Ingeniería se sienta con producto, data y operaciones.

Qué NO es este rol

  • Investigación o fine-tuning: construimos sobre modelos de frontera.
  • Demos: what counts is what holds up in production, measured.

Proceso de selección

  1. Revisión de CV
  2. Entrevista con People (15 min)
  3. Entrevista con Producto (45 min)
  4. Entrevista técnica y live coding con el equipo (1 hora)
  5. Entrevista Hiring Manager (45 min)
  6. Cierre con Founders (30 min)

We look after people's time: we give you feedback at every stage and aim to close the process in a few weeks.

Apply directly on RemoteJobs.org: https://remotejobs.org/remote-jobs/ai-engineer-koltin

Empleos proporcionados por RemoteJobs.org · remotejobs.org

Acerca de este empleo

Esta página describe un empleo en activo: AI Engineer en Koltin, publicado como México.

Frequently asked questions

¿Dónde está ubicado este empleo?
Este empleo figura como México.
¿Qué tipo de trabajo es?
Es un puesto de Jornada completa · Remoto en Koltin.
¿Cuándo se publicó este empleo?
Este empleo se publicó el 9 nov 2026.
¿Dónde solicito este empleo?
Usa el enlace de solicitud de esta página. Abre la oferta original en RemoteJobs.org, donde presentas tu solicitud; StrideRecruit no recibe solicitudes.
Denunciar esta oferta

Cuéntanos si esta oferta parece incorrecta. Revisamos las denuncias; una denuncia no elimina la oferta por sí sola.

No pedimos tu nombre ni tu correo y no guardamos tu dirección IP con la denuncia.