Senior Software Engineer - Agentic AI

Senior Software Engineer — Agentic AI

Help us build an AI agent that doesn’t live inside another app. It texts, remembers, takes action, and runs on its own dedicated machine.

Arktika is assembling the engineering team behind an ambitious agentic AI platform for a fast-growing US company.

The product connects an always-on cloud orchestrator to hundreds of dedicated Mac Mini nodes. Users interact with their personal AI agent through SMS, WhatsApp, Slack, email, and iMessage—no new app or login required.

Behind the scenes, the agent can reason, call tools, execute code, manage workflows, update CRMs, communicate with leads, and coordinate opportunities across a larger network.

This isn’t a chatbot. It’s an operating layer.

The Stack

Python, FastAPI, TypeScript, Next.js, React Native, PostgreSQL, Docker, AWS, Terraform, GitHub Actions, Twilio, OpenAI, Anthropic, and Cursor.

• Remote: Buenos Aires, Argentina
• Compensation: USD 5000 - USD 7000/ month
• Team: Small, distributed, and highly autonomous

What You’ll Build

During your first six months, you could:

• Scale a cloud orchestrator across hundreds of independent AI nodes.
• Build deployment, monitoring, and recovery systems for a physical Mac Mini fleet.
• Stop prompt-injection attacks from exposing private CRM and customer data.
• Design persistent memory across cloud and local agent environments.
• Build isolated Twilio, messaging, email, and CRM integrations.
• Help a central “Queen” node coordinate deals across the wider network.

These are live engineering problems—not ideas waiting on a roadmap.

What We Care About

We’re not counting how many years you’ve used a particular framework.

We want engineers who:

• Turn messy problems into systems that actually work.
• Think seriously about security, failure states, and observability.
• Move comfortably across product, backend, infrastructure, and integrations.
• Use AI coding tools heavily without outsourcing judgment to them.
• Challenge weak assumptions and communicate decisions clearly.
• Know when to ship, when to simplify, and when not to take shortcuts.

Distributed systems, agent infrastructure, security, AWS, or messaging experience will help. Curiosity and sound technical judgment matter more than matching every keyword.

The Honest Part

The system works, but it isn’t finished.

You’ll inherit existing code, replace temporary architecture, build missing internal tools, and occasionally solve problems involving physical machines that cannot be restarted from an AWS dashboard.

Some processes are still being created. Some decisions will be yours.

If you want everything documented, polished, and predetermined, this won’t be comfortable.

If you want to build one of the stranger—and more interesting—AI systems moving into production, let’s talk.

Apply: [Insert link or email]

This engineering team is being built and supported by Arktika, an applied AI firm helping ambitious companies turn AI concepts into production systems.