GenAI Agent Architect
Mumbai, MH, India2d ago
Free account needed to apply or save. It takes a minute, and we'll bring you right back to this job.
Skills
About the role
Brandtech+ (formerly known as OLIVER+) is a global team of creative thinkers, tech-savvy trendsetters, and production pros specialising in film, CGI, automation, AI, motion design, and digital/print content. We partner with over 300 clients in 40+ countries and counting. Our focus is to connect clients with high-quality solutions, talent and ambitious opportunities worldwide.
As a part of The Brandtech Group, we're at the forefront of leveraging cutting-edge AI technology to revolutionise how we create and deliver work. Our AI solutions enhance efficiency, spark creativity, and drive insightful decision-making, empowering our teams to produce innovative and impactful results.
Role: AI Agent Architect
Location: Remote working until further notice
A Little Bit About the Role:
Step into the vibrant world of our eCommerce Content Factory, where you will be instrumental in shaping how high-volume beauty content is crafted and localised across multiple European markets for the skin and hair categories. In this role, you will design, govern, and continuously improve the AI agent architecture that powers content creation workflows – balancing creativity with compliance, quality, and localisation accuracy. You’ll help the team move from “automation projects” to a scalable, governed multi-agent system that increases production capacity now and builds durable in-house capability over time.
In this role, you will help power high-volume, multi-market content production for beauty categories (skin and hair). The agent architecture you build will support consistent creation and localisation of static and, where relevant, motion-capable assets – ensuring outputs remain compliant with local legal/scientific requirements and meet brand standards.
What you will be doing:
Agent Architecture & System Design
- Define the architecture and behaviour of AI agents across creative, production, and localisation workflows for beauty eCommerce content.
- Design end-to-end agent workflows that translate product inputs (e.g., assets, briefs, product information) into repeatable deliverables for multiple markets.
- Build robust multi-step pipelines that incorporate routing, fallbacks, escalation paths, and human checkpoints.
Governance, Standards & Quality
- Define standards and governance models for how agents operate, interact, and improve over time.
- Implement human-in-the-loop QA processes to ensure output accuracy, compliance, and brand/creative quality.
- Collaborate with prompt engineering and engineering partners to develop prompt frameworks and evaluate agent performance using repeatable metrics.
Collaboration Across Teams
- Partner with Product and Engineering to identify new agent opportunities and prioritise development based on business value and delivery impact.
- Translate operational & domain expertise into scalable decision logic aligned to real-world marketing and production use cases.
- Collaborate with creative & production teams to understand requirements, constraints, and quality expectations across markets.
Documentation & Traceability
- Document workflows, dependencies, guardrails, and performance metrics to ensure consistency and traceability across lifecycles.
- Contribute to the evolution of the Agent Architecture framework so Pencil/Brandtech can scale safely, intelligently, and creatively.
What you need to be great in this role:
- Extensive experience designing and improving automation and AI-driven workflows in an agency, tech, or SaaS environment (ideally within marketing/creative or production operations).
- Strong understanding of end-to-end marketing/production workflows, especially those involving briefing, localisation, format compliance, delivery, and quality control.
- Hands-on experience architecting and developing AI agents, prompt frameworks, and model-driven workflows.
- Proven ability to translate complex creative/production processes into structured agent logic and measurable success criteria.