AI/LLM Engineer: RAG Pipelines & Open-Source Models

Salvo Software LLC; México; Presencial; MXN 1 679 000 - 2 799 000; Tempo integral

Resumo da oferta

Salvo Software is a global technology company with distributed teams across the US, LATAM, and India, expanding its AI department. We are building and operating LLM-powered systems, including open‑source models via Ollama/llama.cpp, RAG pipelines, and MCP integrations that connect LLMs to internal tools and data sources.

You’ll work in a fast-paced environment, leveraging production-grade Python, strong fundamentals, and a penchant for learning quickly.

Qualificações

• 3–5+ years of production-level Python for building services, LLM inference, and automation.
• Hands-on LLM integration with Ollama/llama.cpp and vector retrieval.
• Cloud deployments (Azure/AWS) and containerization experience.

Responsabilidades

• Serve and operate open-source LLMs using Ollama and llama.cpp, locally and on‑prem.
• Build and maintain RAG pipelines: embeddings, vector databases, chunking, retrieval quality.
• Develop MCP integrations connecting LLMs to internal tools and data sources.
• Build Python back-end services powering ML inference and AI features.
• Parse structured/unstructured data (XML/XSD, Office formats) for pipelines.
• Advance model optimization: quantization, GPU tuning, offline/air-gapped deployments.
• Support CI/CD pipelines for AI services; deploy workloads to Azure/AWS and Dockerize services.
• Automate operational tasks with Python, Bash, PowerShell; troubleshoot stack end-to-end.
• Collaborate with engineering/product teams to scale LLM-powered features.

Conhecimentos

Python engineering

LLM integration

RAG pipelines

DevOps fundamentals

Cloud deployments

Ferramentas

Docker

Azure

AWS

Kubernetes

Terraform

GitHub Actions

Ollama

llama.cpp

Kafka

Descrição da oferta de emprego

Salvo Software is a global technology company with distributed teams across the US, LATAM, and India, expanding its AI department. We are building and operating LLM-powered systems, including open‑source models via Ollama/llama.cpp, RAG pipelines, and MCP integrations that connect LLMs to internal tools and data sources.

You’ll work in a fast-paced environment, leveraging production-grade Python, strong fundamentals, and a penchant for learning quickly.