About the role

What you will do

•  Architect and implement a workflow orchestration engine for multi-step, long-running, stateful processes — including state-machine modelling, event-driven execution, and durable workflow state.
• Design and build a policy / authorization engine supporting fine-grained access control (RBAC / ABAC or OPA-style policy-as-code), evaluated deterministically and fail-closed.
• Develop a rule / decision DSL and the evaluation engine behind it, so that governance logic is expressed declaratively, versioned, and testable.
• Guarantee correctness under failure through idempotency, distributed locking, exactly-once / at-least-once semantics where appropriate, retries, and fault recovery.
• Own the API contracts between the backbone and every consuming service — clear, versioned, backward-compatible interface design.
• Build for multi-tenancy and isolation with security, data separation, and tenant-aware policy enforcement as first-class concerns.
• Make the platform auditable with comprehensive, tamper-evident audit logging and decision provenance suitable for regulated, compliance-driven environments.
• Set engineering standards for the backbone — design reviews, testing strategy, observability, and the deterministic / fail-closed patterns the rest of the team builds on.

What we are looking for (required)

• Distributed systems depth 
• Workflow orchestration & state machines
• Event-driven architecture — strong command of asynchronous, message- or event-driven design.
• Correctness primitives — hands-on experience with idempotency, distributed locking, consistency models, and fault / failure recovery.
• API contract design —
• Policy / authorization engines 
• DSL / rule-engine development — you have designed or implemented a domain-specific language or rule/decision engine.
• Security, multi-tenancy & audit 

Technical skills & stack

• Languages — strong command of a systems / backend language such as Go, Rust, or Java / Kotlin, with the maturity to choose the right tool for the job. Python familiarity for working alongside the AI engineering team.
• Workflow orchestration — durable-execution / workflow engines such as Temporal (or Cadence, Camunda / Zeebe, AWS Step Functions); building long-running, stateful, recoverable workflows.
• Policy & authorization — policy-as-code and fine-grained authorization in production: OPA / Rego, OpenFGA or Zanzibar-style systems (e.g. SpiceDB), Cerbos, Oso, or Casbin.
• Rule engines & DSLs — expression / rule evaluation (e.g. CEL, Drools) and parser / interpreter work (ANTLR or hand-built); having designed and shipped a domain-specific rule language.
• Messaging & eventing — event-driven, asynchronous architectures on Kafka, NATS, RabbitMQ, or Pulsar.
• Data & state — PostgreSQL with strong transactional and consistency modelling; event-sourcing / append-only stores; Redis. Distributed SQL (CockroachDB / Spanner-style) a plus.
• Coordination & locking — distributed locks, leader election, and coordination using Redis, etcd, ZooKeeper, or Consul.
• Infrastructure & deployment — Docker, Kubernetes, and Terraform; operating services on a major cloud (AWS, GCP, or Azure).
• Observability & operations — OpenTelemetry, Prometheus, Grafana, and distributed tracing; you instrument and run what you build.
• Security & audit — secrets management (e.g. Vault), encryption in transit and at rest, tamper-evident / append-only audit logging, and compliance-aware design.
• Testing & correctness — contract testing, property-based testing, and deterministic / simulation testing for distributed systems.

Strongly preferred

• Model-governance patterns — familiarity with governing, gating, or controlling AI/ML model behaviour in production.
• Deterministic execution — experience building systems where reproducibility and deterministic outcomes are a hard requirement.
• Fail-closed safety mechanisms — you default to safe failure: when in doubt, the system denies rather than guesses.
• Regulated-domain experience — prior work in healthcare, fintech, aerospace, defense, or another compliance-heavy industry.