About the job
Who We Are Looking For
Creating an AI agent is straightforward. The real challenge lies in ensuring it delivers consistent and accurate responses multiple times, particularly when dealing with real enterprise data in a regulated setting, without human oversight on every output.
You have experience deploying AI agent systems in production environments with messy, unstructured data, not just on sanitized demo datasets. You understand the pitfalls of accuracy and know that simply refining prompts won't resolve deeper semantic issues. Your expertise includes building systems that are resilient, even when the model doesn't perform flawlessly every time, and you possess the judgment to know when to proceed with deployment.
You are not seeking a predefined architecture to implement; rather, you are eager to tackle the unsolved challenges and create the solutions that define industry standards.
About Flipside Crypto
Flipside Crypto develops AI solutions that transform chaotic institutional data into actionable insights, workflows, and outcomes. With a strong foundation in blockchain data infrastructure, spanning 8 years, over 20 blockchain networks, and more than 700 million resolved wallets, we now extend our capabilities to enterprises facing similar challenges: leveraging their data efficiently at scale without the need for extensive analyst teams.
Our technology is already in active use by financial institutions and organizations like Interlochen, demonstrating a robust architecture. We are now focused on building the enterprise operations around our proven solutions.
The Role
In this position, you will take ownership of the architecture that ensures the reliability of our AI agent fleets. This includes the harness, tooling, orchestration patterns, and semantic layers that maintain outputs within the context of the organization. Your work will involve actively developing and enhancing Forge (our agent framework), Lattice (fleet orchestration), and Stratum (semantic intelligence), the systems that underpin our production deployments.
Your commitment to output quality is unwavering, not due to external pressure, but because you have witnessed the consequences of agents losing their effectiveness. You will strive for quasi-determinism, utilizing validated tools to ensure agents produce consistent and auditable results at scale. This is a Staff-level role where you will define the architecture, establish standards, and make informed decisions independently.
Your Responsibilities
Design and construct the architecture for AI agent workflows, including planning loops, tool utilization, memory management, retrieval processes, and human-in-the-loop checkpoints.
Assess, integrate, and optimize foundation models and LLM APIs tailored for specific enterprise applications.
