Systems Engineer, Linux and AI

Visa · Fintech · Ashburn, VA

This role involves working with AI and Linux systems, focusing on building and deploying agentic workflows, traditional ML models, and deep learning/LLMs. The engineer will design, build, and maintain agentic workflows, deploy models as APIs, and integrate them with existing applications. Responsibilities include data preparation, model training, evaluation, and implementing orchestration around tools and plugins, as well as supporting Linux systems.

What you'd actually do

  1. Work with stakeholders to translate business problems into AI use cases, define requirements, and design solution architectures.​
  2. Collect, clean, and prepare data, build and train models (traditional ML and deep learning/LLMs), evaluate performance and iterate.​
  3. Design, build, and maintain agentic workflows where LLM “agents” can plan, call tools, and coordinate multi‑step tasks (e.g., retrieval, calling APIs, writing to databases).
  4. Deploy models as APIs or services, integrate with existing applications, implement and tune orchestration around tools and plugins (MCP‑style connectors, function‑calling schemas, vector DBs), including prompt design, safety/guardrails, and evaluation of agent behavior. Ability to isolate and resolve critical system issues to minimize impact and downtime, strong analytical judgment under minimal supervision.

Skills

Required

  • Python
  • shell scripting
  • Linux system administration
  • large language models and GenAI
  • prompt engineering
  • retrieval-augmented generation
  • evaluation of LLM-based applications
  • agentic/LLM orchestration frameworks
  • tool-calling ecosystems

Nice to have

  • Project Management certification
  • ITIL Foundation
  • Git
  • XML
  • JSON
  • XSLT
  • encryption/security technologies
  • SSL
  • IPsec
  • data structures
  • algorithms
  • statistics
  • SQL
  • ServiceNow
  • Red Hat certification
  • RHCSA
  • RHCE
  • Vendor coordination
  • remote install support

What the JD emphasized

  • agentic workflows
  • LLM agents
  • multi-step tasks
  • deploy models as APIs or services
  • orchestration around tools and plugins
  • prompt design
  • safety/guardrails
  • evaluation of agent behavior
  • critical system issues
  • Linux system administration
  • large language models and GenAI
  • prompt engineering
  • retrieval-augmented generation
  • evaluation of LLM-based applications
  • agentic/LLM orchestration frameworks
  • tool-calling ecosystems

Other signals

  • design, build, and maintain agentic workflows
  • deploy models as APIs or services
  • build and train models (traditional ML and deep learning/LLMs)