We are looking for a senior AI Agent Engineer to build agents that execute real tasks, not just answer questions: triage, classification, operations automation, integration with WhatsApp, email, CRM, and internal APIs.
You will design agent behavior, define when it acts alone and when it hands off to a human, and build guardrails that prevent it from taking the wrong action in production.
We expect someone who treats autonomous agents as critical systems: with testing, monitoring, and clear boundaries of responsibility, not as unsupervised magic automation.
Build AI agents that execute real tasks in production, safely and with oversight.
Responsibilities
- Design and implement AI agents for triage, classification, and automation tasks.
- Define the tools an agent can call and its boundaries of action.
- Implement human handoff for cases requiring judgment outside the agent's scope.
- Integrate agents with channels like WhatsApp, email, CRM, and internal APIs.
- Build security guardrails to prevent improper agent action.
- Monitor agent behavior in production and adjust based on real evidence.
- Evaluate agent quality and accuracy with structured evals before and after go-live.
Requirements
- Solid software engineering experience with a focus on building AI agents.
- Command of agent orchestration frameworks, tool calling, and RAG.
- Knowledge of system architecture to safely integrate agents into existing products and channels.
- Hands-on experience with cloud environments to host and scale agents in production.
- Rigor in testing and evaluating agent behavior before granting real autonomy.
- Clear communication to define with the business where the agent can and cannot act alone.
Nice to have
- Experience with multiple agent frameworks (LangGraph, CrewAI, or equivalents).
- Background integrating agents with messaging systems (WhatsApp Business API).
- Knowledge of security specific to autonomous agents (prompt injection, unintended actions).
- Experience with observability and tracing of agent decisions.
- Publications or personal projects with AI agents in production.
- Python
- LangGraph
- Anthropic API
- OpenAI API
- AWS
- Remote, flexible setup: you organize your routine around deliverables, not the clock.
- Growth budget for courses, certifications, and technical events.
- Health and wellness stipend.
- Real access to and ownership of the tools and AI stack we use in day-to-day engineering.
- Results-oriented culture: deliver well, and you get the autonomy to decide how and when you work.
- Small, senior squads with real room for technical decision-making.