Skip to content
View daletoniris's full-sized avatar
👽
Shipeo Cosas...
👽
Shipeo Cosas...

Organizations

@NiperiaLab

Block or report daletoniris

Block user

Prevent this user from interacting with your repositories and sending you notifications. Learn more about blocking users.

You must be logged in to block users.

Maximum 250 characters. Please don’t include any personal information such as legal names or email addresses. Markdown is supported. This note will only be visible to you.
Report abuse

Contact GitHub support about this user’s behavior. Learn more about reporting abuse.

Report abuse
daletoniris/README.md

en es

Hey, I'm Daniel 🧉

The defective bit at the end of the world 🌍💀

📍 Puerto Madryn, Patagonia Argentina — where the whales breach and the packets flow.

Cybersecurity researcher, AI builder, and the kind of person who trains neural networks to detect fire and then builds a WAF to detect attackers. I don't talk about AI — I ship it.

Website Hub


🧬 The Stack That Runs in My Veins

Python TensorFlow OpenCV PostgreSQL Docker Kafka Nginx GCP Linux Claude GPT Raspberry Pi Burp Suite Metasploit


🔥 What I'm Building Now

📖 Alchemind — The Complete AI Journey

"From the first artificial neurons to the quantum consciousness of intelligence."

→ Read the Book + Labs

A comprehensive book, course, and lab environment covering the full arc of AI — from McCulloch & Pitts' mathematical neuron (1943) to autonomous defensive systems deployed in production today. 8 chapters, 2 papers, 4 hands-on labs, and The Warden autonomous defense architecture.

Includes:

  • 📚 8 Chapters — AI history, neural networks, ML algorithms, biases & validation, The Warden, static vs adaptive defense
  • 📄 2 Papers — "Redes Neuronales como nunca te las explicaron" (2018), "La Realidad que Aprende" manifesto
  • 🧪 4 Labs — Classification, threshold tuning, confusion matrices, active learning pipelines
  • 🏗️ The Warden — Autonomous defense architecture: KNN + LLM + active learning feedback loop

Python scikit-learn TF-IDF KNN Active Learning Claude/GPT


🛡️ TokioAI — Autonomous SOC Platform

"An AI that protects you while you sleep."

→ Source Code | → Mirror on my profile | tokioia.com

Not a dashboard. Not a monitoring tool. A full Security Operations Center that thinks, decides, and acts autonomously.

  • 🧠 26 WAF signatures + behavioral rules + honeypot traps + IP reputation scoring
  • Instant blocking — critical attacks die on the first request
  • 🔗 Multi-phase correlation — tracks attackers across recon → probe → exploit → exfiltration
  • 🤖 AI Agent (30+ tools) — Claude, GPT-4, Gemini with automatic fallback, 3 retries + exponential backoff
  • 🔧 Integrations — Home Assistant (LocalTuya), Alexa, IoT devices, Kafka, Tailscale mesh, Telegram
  • 📊 Document generation — PDF reports, PPTX presentations, CSV exports via natural language
  • 🚁 NEW: Drone Control — DJI Tello flight control via natural language: takeoff, patrol patterns (square, triangle, circle, zigzag, sweep), camera, telemetry, and full simulation mode without hardware

Python FastAPI PostgreSQL Kafka Nginx Docker GCP Anthropic Claude OpenAI GPT-4 Google Gemini DJI Tello

🧠 AI Resilience Hub — Ekoparty AI Village

airesiliencehub.com

We lead the AI Village at Ekoparty, the largest security conference in Latin America. Research, workshops, and live demos on AI-powered cybersecurity — offensive and defensive. Building the bridge between AI and infosec communities.


✍️ Niperia Lab — Writings

"Hacking is breaking — code, narratives, ego, illusions. What remains is real."

→ Read the Writings

78 texts extracted from @niperialab Instagram stories — poetry, essays, aphorisms, and reflections written between 2022 and 2026. Where engineering meets philosophy: Jungian shadow, cyclic civilizations, consciousness, technology, and the art of breaking your own preconceptions.

  • 🌙 Poetry & Lyric Texts — Verses on existence, growth, and the inner search
  • 🔮 Consciousness, Ego & Shadow — 25 explorations on Jungian shadow, integration, and hidden patterns
  • 🤖 Philosophy of Technology — AI, biological vs artificial processors, the future of consciousness
  • 📜 Essays & Reflections — Long-form thoughts on life, process, and personal evolution

Philosophy Psychology Jung Technology Poetry Consciousness


🗺️ The Journey — Every Project Led Here

I didn't start in cybersecurity. I didn't start with code. I started as a kid in my brother's electronics workshop in Villaguay, Entre Ríos — a small neighborhood facing a cemetery. We built machines with our hands: sound devices, mosquito repellers, FM radio transmitters, spy microphones. Then I built scientific instruments from nothing — a pluviometer, a barometer, a hygrometer made from human hair, an ant farm, a telescope from old lenses. That kid never left.

🔌 Childhood — The Workshop Next to the Cemetery

The real beginning. A small town in Villaguay, a big brother, and an electronics workshop facing a cemetery. Building machines from scraps: FM transmitters, spy mics, mosquito repellers. But also obsessed with nature — hours observing insects and plants, trying to understand how they worked. Built an ant farm to watch emergent behavior, a telescope from old lenses to map the sky, scientific instruments from nothing (pluviometer, barometer, hygrometer made from human hair). Once dissected a frog with a Gillette razor blade following a manual — because he needed to know how things work inside. That impulse never left: the tool changed from a Gillette to a debugger to TensorFlow, but the drive is the same.

🔓 2007-2012 — The Return & Cybersecurity Origins

At 25, the curious kid came back. Started with malware reversing, migrated to Chubut, Patagonia. From 2010-2012, reported multiple vulnerabilities that improved security of important provincial systems — real responsible disclosure on production infrastructure. Founded Chubut Hack to bring hacking culture to a region where none existed.

🤖 2012-2014 — Robotics, Drones & Hardware

Back to building physical machines. The same impulse from the workshop, now with better tools.

Project What I Built
🚁 drone-hacker-shield Arduino flight controller shield for quadcopters. MPU6050 sensor fusion, PID stabilization, motor mixing. Building drones from scratch.
🛩️ drone-pilot Autonomous pilot for companion computers. MultiWii/Pixhawk control, PID hover, neural network flight controller, OpenCV vision, waypoint navigation.
🎮 fly-drone-controller Python drone toolkit: MultiWii protocol, joystick, waypoints, neural nets, hover controllers, IMU plotting, face detection. The Swiss Army knife of my drone lab.
📡 multiwii-sonar MultiWii firmware + ultrasonic sonar for indoor altitude hold.
📡 multiwii-sonar-v2 Improved sonar integration — better filtering, more stable hover.
🌡️ multiwii-humidity-sensor MultiWii firmware + DHT environmental sensor. Drones that measure their environment.
🧱 rasti-block-robotics Robotics with Rasti Block (Argentine building blocks) + Arduino + IR remote. Innovation doesn't need expensive equipment.

🧠 2015-2017 — AI with TensorFlow & JarvisIA

Project What I Built
🤖 JarvisIA First Spanish voice assistant. Wit.ai NLP + Google Speech + IoT control + Telegram. Built in Puerto Madryn for Latin America. Pioneer project.
👁️ tensorflow-object-detection Real-time detection on Raspberry Pi. Custom training pipeline from scratch.
🌱 deteccion-de-plagas Plant pest detection for 30+ species. MobileNet + Android.

🔬 2018-2020 — Deep Learning & Computer Vision

Project What I Built
🔓 captcha-solver-ml Three different ML approaches to CAPTCHA solving.
🔥 fire-detection-yolov3 Real-time fire detection with custom YOLOv3. Because Patagonia burns.
🛰️ ndvi-drone-analysis NDVI vegetation maps from drone imagery with OpenCV.

🎙️ 2021-2022 — Breaking AI, Offensive Security & Building Community

Project What I Built
🐜 ant-project Digital formicarium — Raspberry Pi + PiCamera + DHT22 + light sensor. Time-lapse recording of ant colonies with environmental data. From childhood ant farms to digital observation systems.
💥 adversarial-ml-attacks FGSM attacks on MobileNetV2. If you can fool a neural network, you can also defend one. This is where TokioAI was born.
🎤 security-talks Slides from Chubut Hack — the cybersecurity community I founded. 20+ talks, speakers from all over LATAM.
⚔️ htb-writeups HackTheBox writeups. Offensive skills that built my defensive vision.

🛡️ 2024-2026 — Everything Converges

Project What I Built
📖 Alchemind The Book. Complete journey through AI: history, neural networks, The Warden autonomous defense architecture, 4 hands-on labs, 2 papers. Everything I've learned in 13 years, compressed for you.
📄 WAF-AI-Paper Published academic paper: hybrid WAF combining Naive Bayes + GPT. Autonomous dynamic learning. Zero-day detection.
🏰 TokioAI Platform Production autonomous SOC. 26 WAF signatures, 30+ tools, multi-LLM agent, real-time ML classification, Telegram bot, Home Assistant/IoT/Alexa integration.
🕵️ agent-smiths Simple React autonomous security agent. LLM-driven planning, multi-tool orchestration (nmap, testssl, nikto).
🧪 adversarial-waf-ml TensorFlow models that generate adversarial WAF evasion payloads through feedback loops.
🏴 ai-security-ctf 5 CTF challenges for Ekoparty AI Village. Docker-ready.
📡 iot-traffic-monitor KNN + Gemini AI for anomalous IoT traffic detection. Active learning pipeline.
🐳 pentest-docker-tools Dockerized pentesting: remote nmap + AI-assisted Kali interface.
⚛️ quantic-encoder Hybrid WAF log classifier: KNN + DeepSeek LLM. On-demand A100 GPU via Terraform, Stripe SaaS.
🎓 ekoparty-ml-security ML for threat detection — Ekoparty AI Village training materials.

📐 Publications & Papers

Year Title Type
2018 Redes Neuronales como nunca te las explicaron Educational Paper
2024 WAF Enhanced with AI through Autonomous Dynamic Learning Academic Paper
2025 La Realidad que Aprende Manifesto
2026 Alchemind Book + Labs

🧉 Philosophy

It started in a workshop next to a cemetery.

Villaguay, Entre Ríos. A kid and his older brother
building electronic machines from scraps: FM radios,
spy microphones, mosquito repellers.

But also outside — in the dirt, under rocks,
between plants. Watching insects for hours.
Studying how ants organize without a leader.
Building an ant farm to observe emergent behavior.
Dissecting a frog with a Gillette, following a manual,
because he needed to know how things work inside.

A pluviometer from a tin can. A barometer.
A hygrometer made with human hair.
A telescope from old lenses. A stellarium.

That kid didn't have a computer. He had solder,
curiosity, a razor blade, and a brother who
taught him that if you can imagine it,
you can build it.

At 25, that kid came back. Started reverse-engineering
malware. Migrated to Patagonia. Reported vulnerabilities
that fixed real systems. Founded Chubut Hack because
nobody was teaching security in his region.

Then he taught machines to see fire, detect pests,
solve CAPTCHAs. Then he broke them with adversarial
attacks. Then he realized: if you can break AI,
you can build AI that can't be broken.

Puerto Madryn: 45° south latitude, population 120k.
Penguins outnumber developers. The wind teaches
you that persistence isn't optional.

TokioAI is not a startup pitch. It's the convergence
of a lifetime of building things that work, from a
place nobody expects innovation to come from.

"No busco que las máquinas sean humanas.
 Busco que los humanos recuerden cómo aprenden."

"El bit defectuoso del fin del mundo..."

📊 Stats

GitHub stats

Languages


🌐 Connect

tokioia.com · airesiliencehub.space · LinkedIn · Twitter · Telegram


Building autonomous security from Patagonia. 🐧🛡️

If you got this far, you should probably star something.

Pinned Loading

  1. Web-Application-Firewall-Purple-AI-Paper Web-Application-Firewall-Purple-AI-Paper Public

    Web Application Firewall (WAF) Enhanced with AI through Autonomous Dynamic Learning and Generative Models

    Python 16 2

  2. adversarial-ml-attacks adversarial-ml-attacks Public

    FGSM adversarial attacks on neural networks - AI security research

    Python 1

  3. agent-smiths agent-smiths Public

    MCP-based autonomous security agent with LLM-powered analysis, multi-tool orchestration, and automated reporting

    Python 1

  4. quantic-encoder quantic-encoder Public

    Hybrid WAF log classifier: KNN proximity + LLM fallback with active learning. On-demand GPU via Terraform, Stripe SaaS billing, GCS storage.

    Python 1

  5. tensorflow-object-detection tensorflow-object-detection Public

    Real-time object detection on Raspberry Pi with TensorFlow and custom training pipeline

    Python 1

  6. alchemind alchemind Public

    From the first artificial neurons to autonomous defense systems — Book + Labs + Papers on AI history, neural networks, adversarial ML, and The Warden architecture. By Daniel Dieser.

    Python 2