Senior AI Engineer IRC305838

GlobalLogic Inc.; México; Híbrido; MXN 700.000 - 1.100.000; Jornada completa

Descripción de la vacante

GlobalLogic seeks a Senior Software Engineer to design and implement a multi-agent pipeline that automates SDLC workflows across cloud, firmware, and mobile codebases. You will build a supervisor agent to orchestrate specialists and ensure iterative, tested outputs from Jira-like inputs, with a focus on reliability and scalability.

The role emphasizes Python proficiency, MCP strategy, and deep experience with AWS Bedrock and prompt engineering to optimize complex AI-driven workflows across

Formación

• 2+ years building LLM-powered applications that shipped to production.
• Hands-on experience building multi-step agent systems that call tools, handle failures, and iterate.
• Strong Python proficiency.
• Strong understanding of AWS Bedrock ecosystem and knowledge base DB concepts.
• Experience with prompt engineering, RAG, context window management, and function calling patterns.
• Experience evaluating non-deterministic AI systems and using MCP in client/server contexts.

Responsabilidades

• Design and implement a multi-agent pipeline that ingests a ticket and outputs a tested, reviewed code change.
• Build a supervisor/orchestrator agent to manage state, delegate to specialists, and support human-in-the-loop.
• Develop specialist agents for context gathering, codebase understanding, problem analysis, and code review creation.
• Define tool schemas using MCP to interface with ticketing, code repos, build systems, devices, and test frameworks.
• Choose foundation models for stages based on reasoning complexity, latency, and cost.
• Create an evaluation framework with test cases, LLM-as-a-judge scoring, and gates in CI/CD.

Conocimientos

LLM-powered apps

Multi-agent systems

Python

AWS Bedrock

MCP

Prompt engineering

Herramientas

Claude

MCP protocol

Descripción del empleo

Agentic & Multi-Agent Systems, Agentic GenAI, Amazon Simple Storage Service (S3), AWS, AWS Bedrock, DynamoDB (AWS), MCP, Python

The objective of this project is to build an autonomous, multi-agent AI pipeline designed to automate the Software Development Life Cycle (SDLC) for embedded IoT products. The system will operate across diverse environments—including cloud infrastructure, device firmware, and mobile codebases—to seamlessly ingest Jira/ticketing inputs and autonomously drive them through to fully tested, peer-reviewed code changes.

The pipeline relies on a hierarchical, multi-agent orchestration model rather than a single-turn LLM wrapper.

The Orchestrator (Supervisor Agent): Acts as the central brain of the pipeline.

If you have a feeling, even slightly, that this is something for you, don’t hesitate. Our fantastic recruitment team will provide you with more information and guide you through the process

Requirements

Level: Senior Software Engineer

• 2+ years building LLM-powered applications that shipped to production
• Hands-on experience building multi-step agent systems that call tools, handle failures, and iterate, not single-turn LLM wrappers
  Strong Python proficiency
• Strong understanding of AWS Bedrock Ecosystem, and Knowledge Base DB building and maintaining.
• Experience with prompt engineering, RAG, context window management, and function calling / tool-use patterns
• Experience evaluating non-deterministic AI systems (LLM-as-a-judge, regression testing for AI outputs)
• Experience with MCP (Model Context Protocol), client or server side

Models: Cloud-hosted foundation models (Claude)
Tool Protocol: MCP (Model Context Protocol)
Compute: AWS / Linux OS

Infrastructure: CI/CD pipeline knowledge

Preferred:

• Experience with code generation or code understanding agents (coding assistants, automated code review, code search)
• Familiarity with cloud-hosted LLM services (model invocation APIs, guardrails, inference configuration)
• Familiarity with embedded systems or device software development workflows

Job responsibilities

• Design and implement a multi-agent pipeline that takes a ticket as input and produces a tested, reviewed code change as output
• Build a supervisor/orchestrator agent that manages pipeline state, delegates to specialist agents, handles iteration loops (build fails, fix, rebuild), and supports human-in-the-loop interrupts at any stage
• Build specialist agents for context gathering, codebase understanding, problem analysis, requirements generation, code generation, test result analysis, and code review creation
• Design tool schemas using the Model Context Protocol (MCP) that define how agents interact with external systems (ticketing, code repos, build systems, devices, test frameworks)
• Implement a tiered model strategy, selecting the right foundation model for each pipeline stage based on reasoning complexity, latency, and cost
• Build the evaluation framework: curate test cases with domain experts, implement LLM-as-a-judge scoring, and integrate evaluation gates into the deployment pipeline
• Manage context windows and token budgets across multi-step agent workflows that can span hours of execution
• Iterate on agent reasoning quality through prompt engineering, tool-use patterns, failure recovery, and hallucination mitigation

What we offer

Exciting Projects: Come take your place at the forefront of digital transformation! With clients across all industries and sectors, we offer an opportunity to work on market-defining products using the latest technologies.

Collaborative Environment:Expand your skills by collaborating with a diverse team of highly talented people in an open, laidback environment — or even abroad in one of our global centers or client facilities!

Work-Life Balance:GlobalLogic prioritizes work-life balance, which is why we offer flexible work schedules.We offer you the best quality of work life so that you exceed the expectations of our clients, while achieving your professional and personal ambitions.

Professional Development:Our dedicated Learning & Development team regularly organizes English classes, professional certifications, and technical and soft skill trainings. We also offer the chance to travel internationally

Excellent Benefits:We provide our employees with competitive salaries, family medical insurance, extended paternity leave, annual performance bonuses, and referral bonuses.

About GlobalLogic

GlobalLogic, a Hitachi Group Company, is a trusted digital engineering partner to the world’s largest and most forward-thinking companies. Since 2000, we’ve been at the forefront of the digital revolution – helping create some of the most innovative and widely used digital products and experiences. Today we continue to collaborate with clients in transforming businesses and redefining industries through intelligent products, platforms, and services.

Diversity and equal opportunity

We are committed to equal opportunity and welcome applications from all backgrounds. We encourage a diverse and inclusive workplace and value the unique perspectives each individual brings to our team.