Asis Rubi Suarez Architecting & shipping production AI applications, RAG pipelines, and multi-tenant systems.
I build end-to-end AI products with TypeScript, React, Next.js, LLMs, and REST APIs—engineered for production reliability with Human-in-the-Loop (HITL) approval-gated workflows, Clerk multi-tenant RBAC, and account-scoped vector & relational data on Supabase & Neon Postgres.
How I Ship Reliable AI to Production
Moving beyond brittle prompt wrappers: deterministic guardrails, account-scoped data isolation, and approval-gated execution.
// 1. Multi-Tenant Auth & Account-Scoped Context (Clerk + RLS)
const { userId, orgId } = await auth.requireTenantScope(req);
// 2. Grounded Retrieval on Neon Postgres / Supabase (pgvector)
const context = await db.queryAccountContext({
tenantId: orgId,
embedding: await embedQuery(input.prompt),
});
// 3. Structured LLM Inference with Schema Validation
const proposal = await llm.generateStructured({
model: "production-tier",
system: SYSTEM_GUARDRAILS,
context,
schema: ActionProposalSchema,
});
// 4. Human-in-the-Loop (HITL) Approval-Gated API Execution
if (proposal.requiresApproval) {
return await approvalQueue.stageForReview({ orgId, userId, proposal });
}
return await executeScopedAction({ orgId, proposal });Approval-Gated REST APIs
Full-stack React product surfaces paired with approval-gated backend APIs so autonomous LLM actions remain auditable, deterministic, and safe in production.
Clerk + Supabase & Neon Postgres
Strict account-scoped data architecture combining Clerk authentication, Role-Based Access Control (RBAC), and Postgres Row-Level Security (RLS).
TypeScript, React & Vercel
End-to-end type safety from database schema to React UI components, deployed globally on Vercel with streaming responses and low-latency APIs.
Approval-Gated AI Workflow Inspector
Test how Blaidd apps handle tenant isolation, RAG context retrieval, and Human-in-the-Loop approval gates before mutating production databases.
Production Systems & Capabilities
Multi-Tenant AI App Platform
Full-stack AI web applications built in TypeScript and React. Integrates LLM workflows, structured outputs, and account-isolated persistence on Neon Serverless Postgres and Supabase.
Approval-Gated Agentic Workflows
Engineered human-in-the-loop state machines and approval-gated REST APIs in React. Ensures sensitive LLM tool calls and mutations require explicit tenant authorization before execution.
Experimentation & Funnel Systems
High-converting eCommerce and product growth architectures across Shopify and TikTok Shop. Delivered 3x conversion rate lifts and 30%+ sales growth via rigorous A/B testing and CRO.
Experience
AI Engineer / Full-Stack Engineer
Jan 2024 – Present- Ship production AI applications with TypeScript, React, LLMs, and REST APIs deployed on Vercel.
- Architect multi-tenant authentication and RBAC with Clerk; persist account-scoped data on Supabase and Neon Postgres.
- Design and ship full-stack product surfaces in React with approval-gated (Human-in-the-Loop) APIs and strict tenant isolation.
Software Engineer, Growth & Technical Consultant
Jan 2020 – Present- Improved Shopify conversion rates 3x through store architecture, technical CRO, and A/B testing on sales funnels.
- Drove 30%+ TikTok Shop sales growth through funnel optimization, campaign systems, and performance analytics.
- Led remote cross-functional teams on acquisition, retention, and analytics for US and EMEA brands.