Skip to main content
Available for opportunities

Erkin Qarayev

Building production-grade AI systems with multi-agent orchestration, RAG pipelines, and scalable full-stack architectures. 4+ years shipping real software at Technosec.io.

01 / About

Building intelligent systems
that solve real problems.

Software Engineer with 4+ years at Technosec.io, specializing in AI infrastructure, multi-agent orchestration, and production-grade systems. Open-source contributor to the JupyterLab ecosystem.

AI & Multi-Agent Systems

Architected a 15K LoC multi-agent orchestration platform with 14 autonomous agents, LangGraph state machines, and intelligent LLM routing.

RAG & Infrastructure

Built full RAG pipelines with document chunking, embedding generation, vector storage, and semantic search across multiple file formats.

Full-Stack Development

Next.js, React, FastAPI, and TypeScript expertise. Real-time WebSocket streaming, Zustand state management, and clean architecture.

Testing & DevOps

Comprehensive test automation with 80%+ coverage. CI/CD pipelines reducing deployment from hours to under 20 minutes.

0

Years Experience

0

AI Agents Built

0+

Lines of Code

0%+

Test Coverage

0%+

Sprint Completion

0+

Production Incidents Resolved

02 / Experience

Where I've been
building things.

Software Engineer

Technosec.io

2021 - Present
  • Architected multi-agent orchestration platform (~15,000 LoC) with 14 specialized AI agents using LangGraph, LangChain, and FastAPI
  • Built full RAG pipeline with document chunking, Ollama embeddings, vector storage, and semantic search
  • Implemented multi-provider LLM abstraction supporting Ollama, OpenAI, Anthropic Claude, and Google Gemini
  • Developed persistent agent memory system with SQLite-backed pattern learning and confidence scoring
  • Built conflict resolution system for multi-agent collaboration using weighted voting and consensus thresholds
  • Created Next.js 16 / React 19 frontend with real-time WebSocket streaming and agent status visualization
  • Built comprehensive test automation framework — 200+ E2E test cases, increasing coverage from ~40% to 80%+
  • Implemented CI/CD pipelines reducing deployment time from hours to under 20 minutes
  • Built custom JupyterLab extensions with TypeScript and React, reducing repetitive tasks by ~35% for 15+ data scientists
LangGraphFastAPINext.jsReactPythonTypeScriptDockerRAGMulti-Agent

03 / Projects

What I've built.

From multi-agent AI platforms to real-time dashboards, each project represents a deep dive into solving complex engineering challenges.

Featured

Multi-Agent Orchestration Platform

~15,000 Lines of Code | 14 Autonomous Agents

Production-ready platform using LangGraph, LangChain, and FastAPI. Features agents with real autonomy — accept, refuse, question, retry, and delegate decisions with a LangGraph state-machine supervisor and intelligent LLM-based routing with rule-based fallback.

LangGraphLangChainFastAPIPythonMulti-Agent

RAG Pipeline System

Semantic Search & Document Intelligence

Full retrieval-augmented generation pipeline with document chunking, Ollama embedding generation, vector storage, and semantic search — enabling agents to ground responses in uploaded documents with configurable top-K retrieval.

RAGEmbeddingsVector SearchOllama

Real-Time Agent Dashboard

Next.js 16 / React 19 Frontend

Full-featured frontend with real-time WebSocket streaming, chat history management, file upload for RAG, and agent status visualization using Zustand state management and Recharts.

Next.jsReactWebSocketZustandRecharts

CI/CD & Test Automation

200+ E2E Tests | 80%+ Coverage

Comprehensive test automation framework using Selenium and Playwright across Chrome, Firefox, and Safari. Integrated into GitLab CI/CD with parallel execution, reducing QA cycle time from 3 days to under 4 hours.

SeleniumPlaywrightGitLab CI/CDPercyJenkins

JupyterLab Extensions

Open-Source | 90%+ Adoption

Custom JupyterLab extensions built with TypeScript and React, reducing repetitive data analysis tasks by ~35% for 15+ data scientists. Published as open-source with community code reviews.

TypeScriptReactJupyterLabOpen Source

AI-Powered Dev Workflows

Claude Code Integration | MCP Servers

AI-powered development workflows using Claude Code, orchestrating specialized agents for code review, security analysis, and automated testing with custom MCP servers, hooks, and prompt engineering strategies.

Claude APIMCPPrompt EngineeringSecurity

04 / Skills

Technical expertise.

From low-level infrastructure to high-level AI orchestration, a comprehensive toolkit honed through real-world production systems.

AI & LLM Orchestration

Architected a 14-agent platform with autonomous decision-making, conflict resolution, and persistent memory.

LangGraphCore
LangChainCore
Multi-Agent SystemsCore
RAG PipelinesCore
Prompt EngineeringAdvanced
Claude APIProvider
OpenAI APIProvider
OllamaProvider
Embeddings & VectorsData
MCP ProtocolAdvanced

Languages & Frameworks

Full-stack development with Python and TypeScript, building real-time dashboards and FastAPI backends.

PythonPrimary
TypeScriptPrimary
ReactFrontend
Next.jsFrontend
FastAPIBackend
PydanticBackend
JupyterLabTooling

Testing & Quality

Built 200+ E2E tests with 80%+ coverage, reducing QA cycles from 3 days to under 4 hours.

SeleniumE2E
PlaywrightE2E
pytestUnit
VitestUnit
Visual RegressionQA
E2E AutomationQA

DevOps & Infrastructure

CI/CD pipelines reducing deployment from hours to minutes, with parallel execution and Docker containerization.

CI/CD PipelinesCore
JenkinsCI
GitLab CI/CDCI
DockerInfra
RESTful APIsAPI
WebSocketAPI

Learning journey

Key milestones in my growth as an engineer.

2021

Joined Technosec.io as Software Engineer

Started building production web apps and test infrastructure.

2022

Scaled test automation to 200+ E2E tests

Achieved 80%+ coverage across Chrome, Firefox, and Safari.

2023

JupyterLab extensions adopted by 90%+ team

Open-source React/TypeScript plugins reducing repetitive tasks by 35%.

2024

Architected multi-agent AI platform

14 autonomous agents, 15K LoC, LangGraph state machines.

2025

AI-powered dev workflows with Claude & MCP

Integrating AI into CI/CD lifecycle with custom MCP servers and hooks.

05 / Contact

Let's build something
extraordinary.

I'm always interested in discussing new opportunities, AI architecture challenges, or innovative engineering problems.