We are looking for a Senior AI Engineer to design and build the intelligence layer of ZyG.
You will own the agentic workflows, LLM integrations, and reasoning pipelines that allow our AI agents to autonomously analyze markets, make decisions, and drive eCom growth at scale.
What you'll do- Design & Build Agentic Workflows: Architect multi-agent pipelines — including planning, memory, tool use, and decision loops — that power autonomous media buying and growth operations.
- Own LLM Integration: Select, prompt-engineer, fine-tune, and evaluate LLMs to produce reliable, high-quality outputs across diverse business tasks.
- Build RAG Systems: Develop retrieval-augmented generation pipelines with vector search and context management to ground agent reasoning in real business data.
- Drive Evaluation & Reliability: Define evals, build testing frameworks, and continuously improve agent output quality, consistency, and safety in production.
- Collaborate Across the Stack: Work closely with backend engineers to integrate AI capabilities into core product APIs, ensuring low-latency, production-grade deployment.
- 8+ years of engineering experience, with at least 2 years focused on LLM-based systems, agents, or applied ML in production.
- Agentic Systems Expertise: Hands-on experience building multi-agent architectures, tool-calling workflows, and orchestration frameworks (e.g. LangGraph, CrewAI, ADK, or custom).
- Prompt Engineering & Evals: You treat prompts as code — versioned, tested, and measured. You know how to systematically debug and improve LLM behavior.
- AI-Native Development: You actively use agentic coding tools (Claude Code, Cursor, etc.) to accelerate your own workflow.
Skills Required
- 8+ years of engineering experience
- At least 2 years focused on LLM-based systems, agents, or applied ML in production
- Hands-on experience building multi-agent architectures, tool-calling workflows, and orchestration frameworks (e.g. LangGraph, CrewAI, ADK, or custom)
- Prompt engineering, versioning, testing, and evaluation of prompts and LLM behavior
- Experience using agentic coding tools (e.g., Claude Code, Cursor) to accelerate development
What We Do
ZyG operates an agentic operating system for e-commerce scale, helping product inventors, entrepreneurs, and brands turn products into successful direct-to-consumer businesses. Its platform uses AI agents, unified data infrastructure, predictive modeling, and financing to assess product-market fit, execute digital growth, replace fragmented tools and workflows, and reduce the financial risk of scaling new brands from validation through customer acquisition and retention.
.png)
