// JavaScript is disabled. The interactive bits (chat, project gallery, contact form) need JS. Reach me directly: jakub@jsle.eu · github.com/jsle97 · linkedin.

Selected projects: LLMHUB · callAI · xdB · xdVision · xdSearch · xdCLI — all built without npm.

Jakub Śledzikowski — JavaScript Master & AI/Prompt Engineer. 14+ years of programming. 10M+ lines of my own code. 6B+ tokens used. Zero npm installs. I build everything from scratch since 2012 — faster, safer, simply better.

Projects

LLMHUB — Advanced web client supporting 100+ AI models (sandbox, web search 1–720 URLs per request, TTS/STT, voice assistant, file and image generation, memory system, proprietary context management).
xdVision — Real-time screen analysis powered by voice AI.
xdB — Lightweight JSON database with relational capabilities.
xdSearch — Multi-phase AI-powered research automation.
callAI — Unified AI API client for 200+ models across 11 providers.
xdCode — Fully autonomous coder with integrated testing.
xdSandbox — Sandbox service for running LLM-generated code.
xdCLI (v2–vN) — Chat-like orchestrator for filesystem, codegen and automation (CLI, web, VPS).
xdPrompts — API and agent toolkit for production system prompts, 95% success rate.
xdSrv, xdFiles, xdToken, xdRender, xdReq and other XD libraries — 30+ custom Node.js libraries built without npm for 4+ years.

Skills

JavaScript (14+ years)
Node.js Ecosystem (9+ years)
AI/ML Integration & Prompt Engineering
Database Design & Optimization
Real-time Systems (WebSockets, SSE)
Computer Vision & Image Processing
System Architecture & Security
Performance Optimization
CI/CD & DevOps Automation

Contact

Email: jakub@jsle.eu · GitHub: github.com/jsle97 · LinkedIn: linkedin.com/in/jsle-eu/

~/portfolio $ whoami

Jakub Śledzikowski JavaScript Master & AI Engineer

14+ years of programming. 10M+ lines of my own code. 6B+ tokens used. Zero npm installs. I've been building everything from scratch since 2012 — because it's faster, safer, and simply better. Every framework is technical debt waiting to explode. I prefer control.

Let's Talk View Work

years: 14+
lines of code: 10M+
tokens used: 6B+
npm installs: 0

whoami.js

// no frameworks, no dependencies
const me = {
  name: 'Jakub Śledzikowski',
  role: 'Senior JS + AI Engineer',
  stack: ['JS', 'Node', 'AI'],
  philosophy: () => 'simpler, faster, more robust',
};

me.ship();
// → 14+ years, 0 npm installs

01 /

About /* me */

I've been writing JavaScript for 14+ years — since 2012 from scratch, without npm, without frameworks, without boilerplate generators. Over four years ago I stopped installing any external package and built my own ecosystem: 30+ custom Node libraries (xdSrv, xdB, xdFiles, xdToken, xdRender, xdReq and more) that power everything I ship.

My philosophy crystallized into three words: simpler, faster, more robust. Every external dependency is a security hole you can't control. Every framework is an abstraction you still have to understand to debug. Vanilla JS works everywhere — and always.

Today I focus on AI and Prompt Engineering. ~6 billion tokens used, production pipelines with a 95% success rate, a personal LLM client (LLMHUB) that nearly replaced my IDE, and a fully autonomous coder (xdCode). If you want something that still works 10 years from now — we should talk.

02 /

Chat with AI

03 /

Selected projects

04 /

My skills

Everything I wanted, I built myself — and that's exactly why I know it inside out. This approach accelerated my learning and gave me knowledge few possess.

Front-End Mastery

HTML5, CSS3 (Flexbox, Grid, Animations, Variables), Vanilla JavaScript (ES2023+). 14+ years writing JS every day.

HTML5
CSS3
JS
WebGL

Back-End Architecture

Node.js (9+ years), own libraries (xdSrv, xdReq, xdFiles, xdToken…), WebSockets, SSE, JWT, Crypto. No npm.

Node
xdSrv
WS
JWT

AI / Prompt Engineering

Multi-model integration (200+ models, 11 providers via callAI), RAG, agents, memory systems, 95% pipeline success rate.

Multi-model
Prompts
RAG
Agents

Prompt Engineering Excellence

"If you want to use AI like everyone else, mediocrely — don't learn from me."

Advanced prompt design, context management by relevance and size (own system), multi-phase research automation — ~6 000 000 000 tokens used in production.

Systems & DevOps

Linux (Debian), Docker, Git, CI/CD, server management, security (OWASP), VPS orchestration, performance optimization.

Linux
Docker
Git
CI/CD

Data & Realtime

Custom JSON database (xdB) with relational capabilities, WebSockets, SSE, computer vision, image processing.

xdB
Realtime
SSE
CV

05 /