Abdulgeni Abdulaziz
Architecting Production RAG Pipelines, Autonomous Multi-Agent Workflows, and High-Throughput Web Applications.
// Ask my AI anything about my work
RAG-backed conversational interface trained directly on Abdulgeni's architecture data & system metrics.
// System Architecture
AI / RAG Engineering
End-to-end vector retrieval pipelines, autonomous multi-agent systems, Model Context Protocol (MCP) servers, and low-latency LLM orchestration.
// rag-pipeline.ts - Vector Retrieval Engine
import { GoogleGenAI } from "@google/genai";
import { ChromaClient } from "chromadb";
export async function queryKernel(prompt: string) {
const ai = new GoogleGenAI({ apiKey: process.env.GEMINI_API_KEY });
const embed = await ai.models.embedContent({
model: "text-embedding-004",
contents: prompt
});
const chroma = new ChromaClient();
const collection = await chroma.getCollection({ name: "kernel_docs" });
const docs = await collection.query({ queryEmbeddings: [embed.values] });
return ai.models.generateContent({
model: "gemini-2.5-flash",
contents: `Context: ${docs}\nQuestion: ${prompt}`
});
}Automation
Event-driven webhooks, n8n orchestrations, GitHub Action bots, and automated email processing pipelines.
// n8n-workflow-webhook.json
{
"event": "pipeline.trigger",
"source": "github_actions",
"actions": ["n8n_sync", "slack_notify"],
"retry_policy": { "max_attempts": 3, "backoff_ms": 1000 }
}Full Stack
Production web applications engineered with Next.js App Router, TypeScript, Prisma ORM, and PostgreSQL.
// app/api/kernel/route.ts - Next.js App Router
import { NextRequest, NextResponse } from "next/server";
export async function POST(req: NextRequest) {
const { prompt } = await req.json();
// Server-side secured API Proxy
const res = await fetch(process.env.CORE_LLM_SERVICE_URL!, {
method: "POST",
headers: { "Authorization": `Bearer ${process.env.SERVICE_SECRET}` },
body: JSON.stringify({ prompt })
});
return NextResponse.json(await res.json());
}Security & Best Practices
OWASP Top 10 mitigation, secure API proxying, type safety, and input sanitization across LLM boundaries.
// security-sanitizer.ts - OWASP Guardrail
export function sanitizeInput(rawInput: string): string {
// Strip harmful script tags & limit payload length
return rawInput
.replace(/<script\b[^<]*(?:(?!<\/script>)<[^<]*)*<\/script>/gi, "")
.trim()
.slice(0, 4096);
}Multilingual Engineering
Fluent global communication for international remote teams & cross-border AI localization.
Kernel Infrastructure Monitor
Real-time node status, latency benchmarks, and operational metrics powering my full-stack & AI architectures.
Cluster ID: kernel-prod-eu-01 • Continuous Ping Active
95th Percentile: 24ms
Zero Unplanned Downtime
Peak: 5,200 RPS
1,490 Attacks Blocked
Real-time Ping Telemetry (ms)
Live 20-tick sliding window response time
Regional Nodes
Active Subsystem Clusters
4/4 Nodes HealthyRAG Vector Search Cluster
Embedding & ChromaDB
High-dimensional embedding retrieval & similarity matching engine.
Gemini 2.5 LLM Gateway
AI Inference Proxy
Streaming token proxy with OWASP prompt-injection guardrails.
n8n Workflow Engine
Automation Pipelines
Event-driven webhooks, GitHub Actions, and asynchronous execution queues.
Cloudflare Edge CDN
App Router & Cache
Low-latency static asset distribution and Edge API route handler.
// Deployed Systems
3D Skill Matrix20 Technologies
Explore Technology Nodes
Hover over any 3D node to view technology specs and matching system architectures, or click to isolate projects.
Agentic RAG
Autonomous multi-agent research pipeline that formulates queries, filters vector context, and verifies factual accuracy.
Complex multi-step technical research required manual iterative web and document searching across dozens of sources, taking hours per topic.
Built a multi-agent routing architecture using LangChain and ChromaDB that autonomously formulates sub-queries, evaluates document relevance, synthesizes context, and iteratively refines answers.
Cut down manual search iterations by over 80% while maintaining sub-500ms retrieval query speed with strict hallucination checks.
Interactive milestone progression from architecture concept to production scale
Phase 3: Production Stage(Week 4)
Optimized retrieval pipeline to sub-500ms SLA, deployed live web application, and established evaluation benchmarks.
SaaS AI Platform
Production-ready AI micro-SaaS boilerplate with subscription billing, serverless AI proxies, and usage metering.
AI WhatsApp Support Bot
Conversational customer assistant for WhatsApp with product RAG knowledge bases and agent handoff.
Embeddable AI Chatbot Widget
Drop-in website assistant with shadow DOM styling, streaming responses, and customizable design themes.
Agentic PR Review Bot
Autonomous GitHub Action that analyzes PR diffs, detects security vulnerabilities, and posts inline suggestions.
Invoice Extractor MCP Server
Model Context Protocol server enabling Claude and AI agents to extract structured data from paper invoices.
AI Email Automation
Intelligent inbox workflow that categorizes incoming client mail, drafts replies, and syncs calendar schedules.
VulnAlert Bot
Real-time security vulnerability notification system forwarding GitHub Dependabot events to Slack.
GitHub Readme Stats
Dynamic real-time GitHub activity cards, star tallies, commit velocity, and language breakdown for @Abdulgeni.
GitHub Readme Dynamic Cards
[](https://github.com/Abdulgeni) [](https://github.com/Abdulgeni)
Most-Used Languages
Repository language distribution computed across 35+ public and commercial projects:
// Engineering Log
AI Systems Engineer
Full Stack Software Engineer
// Testimonials & System Impact
Direct feedback from engineering leaders and product teams who deployed these AI and full-stack solutions.
“Abdulgeni built our production RAG pipeline from scratch. The hybrid search optimization reduced document synthesis latency from 4.2 seconds down to 320ms while maintaining remarkable accuracy across complex technical documentation.”
Elena Rostova
VERIFIED CLIENTHead of AI Infrastructure•Nexus Automation Labs
// Certifications & Education
Industry Certifications
Formal Education
// Contact & Availability
Abdulgeni Abdulaziz
Full Stack AI Engineer
Open for commercial AI projects, RAG architecture consultations, agent workflow automations, or full-stack software development.