Oracle AI Agents Engineer (Mid / Senior / Lead) - Part Time

Company Description

U2XAI builds hallucination-proof AI Cobots that enhance both Oracle Cloud and On-Prem ERP systems. Our Cobots utilize an enterprise knowledge graph to unify data, workflows, and policies, delivering context-aware, reliable intelligence across Procurement, Supply Chain, Demand Planning, and Logistics. Unlike traditional ERP tools, U2XAI focuses on optimizing business outcomes by improving spend management, balancing supply and demand, mitigating risk, and streamlining logistics.

Role Description

This is a part-time remote role for an Oracle AI Agents Engineer (Mid / Senior / Lead). The engineer will be responsible for developing and implementing AI agents, collaborating with cross-functional teams, analyzing business requirements, and integrating AI solutions within Oracle ERP systems. The role also includes testing and debugging the AI algorithms to ensure robustness and accuracy.

What you’ll do Design agentic workflows for common Oracle processes (P2P, O2C, FP&A, inventory, supplier onboarding, HR ops), including tool selection, guardrails, and escalation paths.

Implement RAG pipelines over Oracle data (Fusion REST, BI Publisher, OTBI, Autonomous DB, Object Storage) with vector retrieval, policy filters, and lineage.

Build OCI-native services (Functions, API Gateway, Streaming, Data Flow) to orchestrate agent tools; package and deploy via DevOps pipelines (Terraform / Resource Manager).

Integrate securely with Oracle Integration Cloud (OIC), Fusion SaaS REST / SOAP, and event frameworks; handle idempotency, retries, and compensating actions.

Ship user experiences (Oracle Digital Assistant / VBCS / React) for chat, copilots, and agent control centers, including human-in-the-loop approvals.

Instrument & govern : define KPIs, add telemetry, eval harnesses, prompt tests, A / Bs, red-teaming, PII masking, RBAC, and audit trails.

Partner with SMEs (Finance, Supply Chain, HR) to convert policies / SOPs into deterministic tools + LLM policies.

Production ownership : SLOs, cost controls, model / version rollouts, prompt / config registries, and incident playbooks.

Must-have qualifications 4–10+ years building on Oracle Fusion / OCI or adjacent Oracle stack, with hands-on Fusion Cloud REST / BI / OTBI integrations.

2+ years with LLM apps / agents (tool use, function calling, planning, memory, guardrails) in production.

Fluency with Python / TypeScript and at least one framework for LLM apps (LangChain / LangGraph, Guidance, custom orchestration).

Experience building OCI services (Functions, API Gateway, Object Storage, Autonomous DB, Logging / Monitoring).

Solid knowledge of RAG (chunking, embeddings, vector stores), evaluation (hallucination / factuality), and prompt engineering .

Security-by-design : OAuth2 / SAML, Oracle IDCS , least-privilege, secrets management, auditability, GDPR / SOC2 considerations.

Strong systems thinking : workflow design, error handling, retries, backoffs, and observability .

Nice to have Oracle Digital Assistant / GenAI Service, OIC adapters, VBCS, Oracle JET.

Knowledge of procurement, inventory, planning, finance processes.

Event-driven patterns (Streaming, Events), BPMN, decision tables.

Vector DBs (PGVector / Autonomous JSON / Elastic / OpenSearch), Redis.

Model ops : prompt / version registries, eval suites, tracing (OpenTelemetry), guardrails (regex / DSL / policy engines).

Experience with multi-agent systems and planning (ReAct, TAMPER, graph / plan-based agents).

Azure / AWS GAI familiarity for hybrid setups.