Artificial Intelligence Engineer

**About OneByZero

OneByZero is a Frontier Systems Integrator building agentic AI systems for leading banks, telcos, insurers, and retailers across Asia Pacific. We work on mission-critical problems for organizations operating the region's financial and digital infrastructure. Operating exclusively on AWS, we combine deep technical expertise with strong partnership programs that enable innovation, scale, and delivery excellence. We build with people who are technically serious, outcome-focused, and passionate about solving complex enterprise challenges.

**The Role

The Forward Deployed Engineer (Agentic AI) builds production agentic systems inside regulated enterprises, on client infrastructure, with client data, under the constraints that banks, telcos, and insurers actually operate under.

You are embedded in a client squad. You design, build, evaluate, and own production agentic systems end to end. You stay until the system works, the evaluations hold, and the client's team understands what they have. You leave behind something the next engineer can build on.

At OBZ, you work alongside AI coworkers as your development partner. You direct them, review their output, validate it against your evaluation criteria, and use them to move faster than any individual engineer could alone. This is not a novelty. It is the model.

This role is open to exceptional engineers with 4- 6 years of experience. What matters is not years of experience. It is the quality of your thinking, the depth of your curiosity, and the evidence that you have already been building with agentic tools because you wanted to, not because someone asked you to.

**What You Will Do

• Build production-grade agentic systems on AWS and OBZ's Neo platform: multi-agent orchestration, RAG pipelines, tool design, MCP server development, human-in-the-loop workflows, and the integration layer that connects agents to real enterprise systems with real data quality constraints and real security perimeters.

• Write evaluations before you write features. Define what good performance looks like, build the harness that measures it, and test against it throughout the build. In regulated environments, systems that have not been evaluated do not reach production.

• Integrate into enterprise reality. Legacy APIs, inconsistent data availability, access control constraints, and compliance requirements that were not designed for agentic systems. Build within those realities, not around them.

• Review and own what you build through production. Clean code, tested code, documented code. If the system degrades in production, you investigate. If the architecture needs to change, you contribute to the decision. You are accountable for what you ship.