Agentic AI Engineer
New
C
CodeRoadSoftware Development
Latin AmericaFull-TimeMiddle
Salary not disclosed
Apply NowOpens the employer's application page
Job Details
- Languages
- Advanced English (written and spoken) is mandatory.
- Experience
- 3+ years of professional software development experience
- Required Skills
- AWSPython
Requirements
- 3+ years of professional software development experience.
- Direct, hands-on experience building LLM-based, RAG, or agentic AI applications.
- Advanced proficiency in Python.
- Experience with AWS (Amazon Bedrock).
- Proficiency with agentic frameworks such as LangGraph, CrewAI, or ADK.
- Experience with vector databases such as Pinecone or Weaviate.
- Knowledge of Model Context Protocol (MCP).
- Hands-on experience testing, evaluating, and monitoring LLM execution using tools like LangSmith, Langtrace, or AgentOps.
- Advanced English (written and spoken) proficiency.
- Ownership mindset and strong troubleshooting capabilities.
Responsibilities
- Develop production-grade AI agents, state management mechanisms, and multi-agent workflows using frameworks such as LangGraph, CrewAI, or Agent Development Kit (ADK).
- Connect AI agents to enterprise tools, APIs, and external platforms via Model Context Protocol (MCP) and structured function calling.
- Implement and maintain high-performance RAG pipelines using vector databases like Pinecone or Weaviate, applying semantic search, chunking, and grounding strategies.
- Execute automated evaluations, regression testing, task-success metrics, and safety guardrails to defend against prompt injection and sensitive-data exposure.
- Deploy LLM applications on AWS utilizing Amazon Bedrock, while leveraging observability platforms such as LangSmith, Langtrace, or AgentOps to troubleshoot failures and optimize performance.
- Write clean, maintainable, and testable Python code, maintaining comprehensive technical documentation while collaborating with senior engineers, MLOps, and product teams.
View Full Description & ApplyYou'll be redirected to the employer's site
