OpenAI Architect (FDE)

Role: OpenAI Architect (FDE)

Role Summary

Lead the architecture and productionisation of OpenAI‑first solutions in a forward‑deployed model. You will embed with customers to design secure, scalable patterns around ChatGPT Enterprise rollout and administration (SSO/SCIM/RBAC, data controls), OpenAI API endpoints (Assistants & tool/function calling, Responses/Chat Completions, Embeddings, Files/Batch, Moderations), fine‑tuning pipelines, and agentic RAG then drive PoC → Production with governance, observability, and cost control. Keep solutions portable with pragmatic use of cloud services, LangChain/LangGraph/Semantic Kernel, and standard vector stores.

What you’ll do

OpenAI delivery

• ChatGPT Enterprise deployment & governance: Plan workspaces; implement SSO/SCIM, role models and policy guardrails; set up usage analytics; define custom GPTs governance (actions/connectors, approvals) and runbooks.

• OpenAI API architecture: Design patterns for Assistants with multi‑tool orchestration, structured outputs (JSON schemas), function/tool calling, Files/Batch for bulk jobs, Moderations, and Embeddings for retrieval.

• RAG & evaluation: Stand up OpenAI‑centric RAG (chunking, embeddings, indexing), implement groundedness checks, prompt test suites, red‑teaming, and cost/perf SLOs.

• Fine‑tuning lifecycle: Own dataset curation, training/eval, bias checks, rollback/versioning, and telemetry for tuned models.

• Operability: Add observability (OpenAI Observability/OpenTelemetry), token/cost telemetry, retries/backoff, idempotency, and feature flags/canaries; document runbooks and SOPs.

Cross‑platform & enterprise integration

• Azure, and/or AWS or GCP& identity/networking: Design with Managed Identity, Secret Management, and optional Private Link/private endpoints; harden per enterprise controls.

• Frameworks & vector stores: Apply OpenAI Agent SDK, LangChain/LangGraph or Semantic Kernel where useful; integrate Azure Cognitive Search, Redis/pgvector, or managed vector services.

• Copilot/Graph & app embeds: Where valuable, integrate via Copilot Studio and Microsoft Graph; wire assistants into Teams/SharePoint and line‑of‑business apps.

• Delivery engineering: Enforce CI/CD (GitHub Actions/Azure DevOps) and IaC (Terraform/Bicep); support multi‑region rollout strategies.

Minimum Qualifications

• 7+ years in software/solution architecture; strong Python plus one of Java/TypeScript.

• Proven delivery on OpenAI/Azure OpenAI (Assistants/tool calling, RAG, eval/safety, observability) and enterprise deployments (auth, policies, cost).

• Hands‑on CI/CD and IaC; excellent customer‑facing communication.

Preferred (mix‑and‑match)

• ChatGPT Enterprise administration (SSO/SCIM/RBAC), custom GPTs governance, usage analytics.

• Fine‑tuning (dataset QA, training/eval pipelines, regression testing).

• RAG stacks (Azure Cognitive Search, Redis/pgvector), OpenAI Agent SDK, LangChain/LangGraph or Semantic Kernel;

• OpenAI Observability, OpenTelemetry.

• Professional-level cloud certifications

Representative problems you’ll work on

Clinical advisors, shopping assistants, insurance document copilots, ServiceNow analytics, SAP copilots, and utilities/regulatory assistants delivered with agentic orchestration, guardrails, and observability across OpenAI‑first and multi‑vendor contexts.

About HCLTech AI & Cloud Native Labs

We are HCLTech’s global Centre of Excellence guiding advanced‑tech adoption for the world’s largest enterprises—combining strategic advice with accelerated engineering and open‑source leadership (CNCF).