State Machine Engineer

About Us At Metaprise, we’re building the world’s first Financial Operating System — an AI Agent–powered platform that unifies Accounts Receivable, Accounts Payable, Contract Management, SmartFlow Automation, and Cash Management into a single intelligent hub. We help enterprises eliminate operational gaps between financial tools, giving them total control over data flow and cash flow. Backed by real customers in the US and global expansion momentum, we’re scaling fast and looking for builders who thrive in high-growth, high-impact environments. Requirements

A. State Machine/Workflow Engine Direction (Strong Priority)

You have done at least one of the following:

Have designed or modified workflow engines (BPM/Orchestrator)

(Such as Activiti/Camunda/Zeebe/Temporal/self-developed engine)

Have built a Task scheduling system (Task Engine/Job Engine)

Designed a plug-in node system (Plugin Architecture)

Has implemented complex business process orchestration (Event-driven/Saga Pattern)

Have worked on stateful logic and compensation mechanisms in banking/clearing/trading systems

Key words: State transition, Saga, idempotence, retry strategy, compensation transaction, message consistency, replay log, lifecycle modeling

B. Distributed Systems & High Concurrency Direction (Can be regarded as a core advantage)

You should at least be familiar with

Distributed consistency, concurrent control, and lock mechanism

Distributed Transactions (TCC/Saga)

High concurrency service (QPS over 100,000)

Message queues, idempotence, delayed tasks, event-driven architecture

Self-developed frameworks or highly complex system components

Key words: Sharding, high reliability, high consistency, distributed scheduling, Event Sourcing, CQRS

C. Semantic Model/Knowledge Graph (suboptimal, but highly desirable)

Not mandatory, but bonus points:

RDF / OWL / SPARQL

Business Object Modeling (Ontology Modeling)

Graph Schema design

Semantic reasoning (Inference Engine

Responsibilities

1. Enterprise-level State Machine Design (FSM/Lifecycle Design

define the business objects (Invoice/Payment/Contract/Flow Instance) life cycle

Design the state Transition Table (Transition Table

Implement Guard, Side Effects, Rollback, and Compensation

Ensure that the process is deterministic, auditable and replayable

2. Workflow Execution Engine (Orchestrator Runtime)

Implement an execution framework similar to Temporal/Camunda

Support the Plugin Nodes system

Handle concurrency, retry, idempotence, and compensation logic

Supports asynchronous execution and event-driven (Webhook/Queue)

Support simultaneous operation of multiple versions of processes (Workflow Versioning)

3. Node execution Lifecycle and plug-in architecture

Logic Nodes (Conditional Judgment, Branching, Merging)

Integration Nodes (API/Bank/ERP)

AI Nodes (LLM plugin nodes, supporting JSON Schema validation + Fallback)

4. High Reliability Execution Model (Tech Safeguards)

Optimistic Lock/Version Vector (to prevent concurrent conflicts)

Anti-Corruption Layer (Agent Output Sanitization)

Async State Pattern (to avoid freezing)

Schema Evolution / Versioning Strategy

5. Collaborate with the semantic model (Ontology)

Work with semantic engineers

Abstract the business objects as Graph Schema

Make process decisions based on semantic layer rules

In the future, SPARQL-based inference will be supported to trigger process migration