AI Developer

GlobalLogic; Argentina; On-site; ARS 600,000 - 1,200,000; Full time

Job summary

GlobalLogic is seeking an engineer to design and ship agentic AI systems. You will write production-grade Python code, build robust data pipelines, and own the full lifecycle from development to deployment.

You'll integrate LLMs via Bedrock or comparable APIs, connect to Snowflake and OpenSearch, and implement multi-agent orchestration with LangGraph, LangChain, and AgentCore. Collaboration with product, sales, and analytics is expected.

Qualifications

• Hands-on experience building agentic AI systems and production-grade Python code.
• Experience with multi-agent orchestration frameworks (LangGraph, LangChain, etc.).
• Experience integrating LLMs via Bedrock/Anthropic or similar.
• Experience with data platforms like Snowflake/OpenSearch.
• Familiarity with MCP and tool-use patterns.
• AWS exposure including IAM, Lambda, API Gateway is a plus.

Responsibilities

• Build multi-agent pipelines using LangGraph and AgentCore.
• Develop prompt pipelines with LangChain.
• Deploy LLM-powered agents on AWS Bedrock.
• Build FastAPI services exposing agentic capabilities.
• Collaborate with CS, Sales, and Analytics to translate needs.

Skills

Python

LangGraph

LangChain

AgentCore

MCP

AWS Bedrock

OpenSearch

Snowflake

Tools

FastAPI

Job description

The ideal candidate will have hands-on experience building and shipping agentic AI systems, with deep proficiency in Python and the modern AI engineering stack:

• Python Expertise: Python is your primary language. You write clean, production-grade code, build robust data pipelines, and are comfortable owning the full lifecycle from
• Agentic Frameworks: Practical experience with LangGraph, LangChain, AgentCore, or comparable multi-agent orchestration frameworks — including agent routing, state management, tool use, memory, and evaluation.
• LLM Integration: Demonstrated experience integrating large language models via AWS Bedrock, Anthropic APIs, or similar — including prompt engineering, context management, and output validation.
• Data Source Integrations: Hands-on experience connecting applications to modern data platforms including Snowflake, Apache Iceberg, OpenSearch, or similar.
• MCP / Tool-Use Patterns: Familiarity with Model Context Protocol (MCP) or equivalent patterns for giving agents structured, governed access to external data systems.
• Cloud Platforms: Hands-on AWS experience, with Bedrock exposure strongly preferred. Familiarity with IAM, Lambda, and API Gateway a plus.

Beyond technical skills, we’re looking for someone who brings:

• Bias for Action: You ship, instrument, and iterate — you don’t wait for perfect requirements.
• Strong Communication: Ability to translate complex agentic system behavior to non-technical stakeholders clearly and concisely.
• Security and Governance Mindset: Awareness of responsible AI practices, data privacy considerations, and the importance of auditability in agentic systems.
• Collaborative Spirit: Comfortable working across functions and levels, from CSMs to the C-suite.

Job Responsibilities

Here’s how you’ll be making an impact:

• Build Multi-Agent Pipelines: Design and implement orchestrated multi-agent systems using LangGraph and AgentCore, including routing logic, evaluation loops, retry mechanisms, and agent specialization patterns.
• Develop with LangChain: Leverage LangChain to build sophisticated prompt pipelines, tool-augmented agents, memory constructs, and retrieval-augmented generation workflows.
• Deploy on AWS Bedrock: Leverage AWS Bedrock to host, invoke, and manage LLM-powered agents at scale, ensuring reliability, cost efficiency, and security compliance.
• Build FastAPI Services: Develop lightweight, production-ready API services that expose agentic capabilities to downstream consumers and orchestration platforms. instrument your work, and evolve solutions based on real usage signals.
• Collaborate Cross-Functionally: Partner closely with Customer Success, Sales, and Analytics stakeholders to translate business requirements into agentic architectures.