AI Agent Architect
Edisyl, identified in the site’s patent disclosure as Flipside Crypto Inc, provides enterprise semantic-data infrastructure and AI-agent systems. It helps organizations encode internal metric definitions and connect data sources so AI systems, pipelines, and intelligence workflows can operate using the organization’s own business context.
Funding history
About edisyl
Edisyl builds a semantic layer for enterprise data that captures organizational definitions, metric logic, and otherwise undocumented knowledge, then binds that context to existing data systems. Its platform comprises Stratum, the semantic intelligence layer; Forge, an agent framework and tools layer; and Lattice, agent-fleet orchestration. The company deploys these capabilities in client environments for use cases including lead scoring, CRM write-back, automated briefings, multi-source data connectivity, and generation and validation of DBT transformation pipelines. Its foundation includes prior blockchain-data infrastructure and entity-resolution work across multiple chains.
Skills
Candidate Availability
Required and preferred rules are kept separate and reflect the wording in the original posting.
About the Role
You will design and build AI-agent workflow architecture, including planning, tool use, memory, retrieval, and human checkpoints. You will integrate and fine-tune models, establish reliability and observability standards, build evaluation tooling, and make documented architectural decisions for production systems.
Requirements
- 6–10 years building production AI or data systems
- Experience with multi-agent architectures
- Strong Python skills
- Experience with agent frameworks such as LangChain, LlamaIndex, or AutoGen
- Experience with RAG architectures and vector databases
- Experience deploying LLM-powered systems in enterprise contexts
- Knowledge of data security, access controls, and audit logging
Responsibilities
- Design and build AI-agent workflow architecture
- Evaluate, integrate, and fine-tune foundation models and LLM APIs
- Define standards for agent reliability, observability, and failure modes
- Translate client deployment learnings into reusable platform components
- Build evaluation harnesses for agent quality, hallucination rates, and task completion
- Make documented architectural decisions
Benefits
- Meaningful early-stage equity
- Remote work
