Lead ML Engineer: LLMOps, Agents & Production AI

SoftServe; Chile; Presencial; CLP 60.000.000 - 120.000.000; Jornada completa

Descripción de la vacante

SoftServe seeks a senior ML engineering leader to set technical direction for ML systems, overseeing production-grade AI pipelines, LLMOps, and multimodal solutions. You will bridge research and real-world impact, guiding cross-functional teams in a fast-paced CoE.

You will mentor ML Engineers, shape strategy for NLP/RAG and agentic AI, and drive best practices across AWS, Azure, and GCP environments.

Formación

• Proven track record deploying production ML systems.
• Expertise in LLMOps, AgentOps, and experiment tracking.
• Experience with cloud AI/ML on AWS, Azure, or GCP.
• Knowledge of multi-agent frameworks LangGraph or CrewAI.
• Leadership or mentoring of ML engineering teams.
• Strong English communication to clients and execs.

Responsabilidades

• Define standards and architecture for end-to-end ML systems.
• Oversee production-ready LLMOps pipelines.
• Collaborate with Data Scientists and engineers for NLP/RAG.
• Advance agentic AI with multi-agent systems.
• Improve ML engineering processes and CI/CD on cloud.
• Mentor ML Engineers and promote MLOps practices.
• Contribute to technical strategy within the CoE.

Conocimientos

Python

LLMOps

AgentOps

MLflow

LangSmith

Langfuse

Weights & Biases

LangGraph

CrewAI

AWS

Azure

GCP

CI/CD/CT

Team leadership

Stakeholder mgmt

English proficiency

Master's degree in CS

Educación

Master's degree in Computer Science

Herramientas

MLflow

LangSmith

Langfuse

Weights & Biases

GitHub Actions

Kubeflow

Descripción del empleo

SoftServe seeks a senior ML engineering leader to set technical direction for ML systems, overseeing production-grade AI pipelines, LLMOps, and multimodal solutions. You will bridge research and real-world impact, guiding cross-functional teams in a fast-paced CoE.

You will mentor ML Engineers, shape strategy for NLP/RAG and agentic AI, and drive best practices across AWS, Azure, and GCP environments.