English

IKIGAI AI WORKSHOP

San Francisco, USA
5-Day Intensive In-Person AI & Cybersecurity Workshop in San Francisco

Hands-on training in AI agents, automation, cybersecurity, and content systems — build real tools you’ll use immediately.

Reserve your spot View schedule
Beginner → Intermediate

IKIGAI AI WORKSHOP

Hands-on AI + Cybersecurity + Automation

  • calendar
    5 Days

    40 hours total instruction

  • location
    Ai Start Academy SF,

    1769 15th St, San Francisco, CA 94103

  • clock
    9:00 AM – 5:00 PM

    Lunch 12–2

  • teacher
    In-person, hands-on

    80% labs

  • dollar
    From $150 Early Bird

  • people
    Limited to 20 participants

    +2 scholarships

Reserve your spot

What you’ll build in 5 Days

Five practical systems built from scratch using the exact tools and methods covered in the workshop.

Day 1

Secure Personal AI Assistant

A secure, local assistant running in an isolated environment — built after understanding why 94% of AI projects fail to capture value.

You’ll leave with:

  • Token economics calculator + architecture decision matrix
  • Hands-on experience with hallucination forensics and prompt injection attacks
  • Working secure AI assistant with 4-layer defense system
  • Security validation checklist tested against real attack vectors
  • Personal AI ROI worksheet with MIT/Harvard productivity benchmarks
  • Foundation for Day 2's cybersecurity deep-dive
Day 2

AI-Powered Data Leak Detection & Response

An automated system for detecting, analyzing, and responding to real-world data leaks — built from Instagram, TikTok, and FBI breach case studies.

You’ll leave with:

  • Automated HaveIBeenPwned monitoring + AI validation system
  • GitHub secret scanner with intelligent leak detection
  • Hands-on: analyze real leaked database dumps with AI
  • Breach response playbook based on Instagram/TikTok/FBI cases
  • Cloud vulnerability checklist (AWS, GCP, Azure)
  • Ready-to-deploy monitoring and alerting scripts
Day 3

Personal Voice Agent (JARVIS)

A voice-enabled AI agent that searches the web, manages calendar, and runs real workflows.

You’ll leave with:

  • RAG-powered chatbot built in 3 hours on n8n (no-code platform)
  • Voice-enabled JARVIS: ElevenLabs + VAPI + n8n integration
  • Multi-agent system: Google Calendar + Gmail + Twilio automation
  • Content workflow: article scanning → LinkedIn post generation
  • Web search + tool calling architecture (Tavily API)
  • Production templates for email automation and scheduling
Day 4

Modern Penetration Testing Toolkit

A complete, hands-on penetration testing workflow used in real security engagements — from OSINT reconnaissance to professional reporting.

You’ll leave with:

  • 8-step pentest framework from reconnaissance to reporting
  • Practical experience: nmap scanning + Burp Suite + Metasploit exploitation
  • OSINT techniques + social engineering simulation results
  • CryEye platform access for AI-enhanced vulnerability validation
  • Professional pentest report template used in real engagements
  • Completed lab simulation: full pentesting lifecycle
Day 5

Real-Time Digital Employee (Voice + Video Agent)

A deployable real-time AI agent that can represent you or your business in live conversations.

You’ll leave with:

  • Working digital employee: voice + face + memory orchestrated in real-time
  • Production-ready system prompt with hallucination guardrails
  • Tool-enabled agent connected to calendar, CRM, and data retrieval workflows
  • 3 founder-tested use cases: 24/7 sales rep, L&D coach, investor screener
  • Security framework: permission scoping, output validation, human-in-the-loop protocols
  • Live deployment on your website with engagement tracking
  • Capstone: your agent tested against objections and edge cases in front of the group

Who this workshop is for

Designed for professionals building, securing, or automating real-world AI systems.

  • Founders & Startup Teams

    Building AI-powered products or internal automation systems.

  • Security Analysts & Pentesters

    Looking to integrate AI into modern defensive and offensive workflows.

  • Engineers & Developers

    Working with APIs, workflows, AI agents, or production systems.

  • Marketing & Content Teams

    Building scalable AI-driven content pipelines and automation.

  • QA / DevOps / DevSecOps

    Responsible for deployment, security, and operational reliability.

  • Anyone serious about AI systems

    Who wants practical, hands-on experience — not theory.

Prerequisites

Beginner → Intermediate

  • Recommended laptop: with at least 16 GB RAM
  • Comfortable with installing software, using CLI, using Git, writing/reading code
  • Slight fear that you aren’t caught up
  • Willingness to learn and experiment

Setup instructions will be sent 7 days before the workshop.

Not sure if this is the right fit for you?

→ Contact us and we’ll help you decide

Workshop Schedule

Daily: 9:00 AM - 5:00 PM (2-hour lunch break: 12:00 PM - 2:00 PM)

Register to Attend

Meet your instructors

Bay Area practitioners building AI systems, automation workflows, and security tooling in real environments.

Pavel
Pavel Spesivtsev
Serial AI-Founder & CTO

GapTrap.ai, Fibonacci 7, Cellery.ai

Serial AI-Founder and CTO @ GapTrap.ai, Fibonacci 7, Cellery.ai. Product Builder who runs production AI agent systems operating autonomously for weeks without human intervention or security incidents. Pavel has developed frameworks for agent governance that balance capability with control. Philosophy: 'Slow down to go fast' - build the harness, quality and security layers first, then add features.
Has provided over 40 different cybersecurity services including red team operations for Fortune 500 companies.
linkedin
instagram
twitter
github
Bogdan
Bogdan Red
Founder & CEO

CQR Company

Senior penetration tester and security researcher with 15+ years of hands-on experience. Bogdan has been at the forefront of offensive security since 2008, discovering and responsibly disclosing hundreds of data breaches. He developed CryEye, an enterprise cybersecurity platform.
Has provided over 40 different cybersecurity services including red team operations for Fortune 500 companies.
linkedin instagram twitter github
Kheir
Kheir Lazaref
Founder & CEO

AI Automation Expert

AI automation expert specializing in N8n workflow orchestration and AI agent development. Kheir builds production-grade automation systems that connect AI models to real business processes, enabling teams to scale without hiring.
Transforms complex workflows into elegant automation solutions.
linkedin instagram twitter github
Max
Max Sapo
Founder & CEO

Happyverse.ai (@Delamax)

A pioneer in Real-Time Interactive AI Agents. Max builds the infrastructure for 'Digital Employees'—video agents that can see, hear, and converse with users in real-time (<500ms latency). He specializes in the architecture of autonomous communication.
Replacing static chatbots with human-like video interfaces for sales, screening, and coaching.
linkedin instagram twitter github
Lex
Lex Mustafin
Founder & CEO

AI Start Academy

Serial Entrepreneur dedicated to democratizing AI education and bringing world-class AI & cybersecurity training to professionals worldwide.
Committed to building a community of AI practitioners who can shape the future.
linkedin instagram twitter github

FAQ

  • 1.Do I need to bring my own laptop?
    Yes, participants should bring a laptop with at least 8GB RAM and 20GB free disk space. Detailed setup instructions will be provided one week before the workshop.
  • 2.Is this workshop suitable for beginners?
    This workshop is not designed for beginners, it provides in-depth diving for more experienced participants. We do cover fundamentals though before moving to advanced topics, and instructors provide personalized support.
  • 3.What if I can't attend all 5 days?
    We strongly recommend attending all 5 days as each day builds on the previous. However, if you must miss a day, we provide recordings and materials for self-study through our skool.com closed community access. You would get FREE LIFETIME ACCESS to our closed skool.com community where we will post all recordings from all upcoming workshops.
  • 4. Are meals included?
    Tea, coffee, and snacks are provided throughout the day. Lunch is not included - there is a 2-hour break where you can explore San Francisco's excellent food scene. We'll provide recommendations for nearby restaurants.You may also bring your own lunch - we have a fridge and a microwave!
  • 5.What's your cancellation policy?
    Full refund if you are unable to attend due to any reasons. Full refund if you are unsatisfied with received knowledge.
  • 6.Can I get a custom workshop for my team?
    Absolutely! We offer customized private workshops tailored to your organization's specific needs, technology stack, and goals. Contact us at hello@aistartacademy.com for pricing and availability.
  • 7.Will I receive a certificate?
    Yes, all participants who complete the workshop receive a certificate of completion including the workshop title, dates, total training hours, and topics covered.
Help Center

Get in touch

phone
+1 347 286-5749
mail
hello@aistartacademy.com
  • The Second Inflection: 88% of companies adopted AI, only 6% capture real value (McKinsey 2025). What separates the two groups.
  • Architecture of Intelligence: tokens, embeddings, attention, pre-training, and why Claude, Gemini and ChatGPT feel different
  • Hands-on: Token economics calculator, embedding space exploration, temperature experiments
  • Why AI Fails: hallucination, context overflow, prompt injection, sycophancy, and how to prevent each
  • Architecture Decision Matrix: when to use prompt engineering, RAG, fine-tuning, or agents
  • Hands-on: Hallucination forensics on your own domain; group architecture design challenge
  • From Prompt to Agent: system prompts, few-shot learning, chain of thought, tool use, building your first agent
  • 60-second MCP teaser (dedicated standalone workshop available separately)
  • Business Decision Framework: model comparison table (OpenAI, Claude, Gemini, Kimi K2.5, GLM 5, Qwens), MIT/Harvard productivity data, personal AI ROI worksheet TEST
  • The OpenClaw disaster: 150K+ GitHub stars, 4,500+ exposed instances, plaintext credentials, zero authentication
  • The Lethal Trifecta: why personal AI assistants become security nightmares (private data + network access + untrusted input)
  • Hands-on: Building the secure foundation in four layers: isolated execution environment, authentication gateway, human-approval confirmation flows, network egress allowlist
  • Adding high-value capabilities: persistent memory, messenger integration, connecting to real-world data sources (specifics adjusted based on pre-workshop survey)
  • Testing your defenses: prompt injection, credential exfiltration, unauthorized network egress. Attack your own system. Fix what breaks.
  • What you take home: working secure AI assistant, security validation checklist, architecture decision matrix, personal AI ROI workshee
  • Bridge to Day 2: your assistant becomes a case study for Bogdan's cybersecurity sessions
  • Cozy space for networking and relaxing: Tea, kava, soft drinks, pizza, fire pit.

AI & LLM Platforms:

  • Claude API (Anthropic)
  • ChatGPT API (OpenAI)
  • Google Gemini API
  • OpenAI Tokenizer
  • Hugging Face Transformers

Development & Execution:

  • Docker (isolated execution environment)
  • Python 3.x
  • Node.js / TypeScript
  • VS Code / Cursor IDE
  • Git / GitHub

Security & Authentication:

  • OAuth 2.0 authentication libraries
  • JWT (JSON Web Tokens)
  • Environment variable management (.env)
  • Network firewall configuration tools
  • SSL/TLS certificates

Agent Framework & Tools:

  • LangChain or LlamaIndex (agent orchestration)
  • MCP (Model Context Protocol) - teaser
  • Custom prompt engineering templates
  • Function calling / tool use libraries

Testing & Validation:

  • Prompt injection test suites
  • Credential scanning tools
  • Network monitoring utilities
  • Security validation scripts

Integration & Capabilities:

  • Telegram Bot API / Discord API / Slack API (messenger integration)
  • Redis or SQLite (persistent memory)
  • REST API clients
  • Webhook handlers

Analysis & Worksheets:

  • Token economics calculator (custom tool)
  • Embedding visualization tools (t-SNE/UMAP)
  • Temperature testing sandbox
  • Personal AI ROI worksheet (Excel/Google Sheets template)
  • The Second Inflection: 88% of companies adopted AI, only 6% capture real value (McKinsey 2025). What separates the two groups.
  • Architecture of Intelligence: tokens, embeddings, attention, pre-training, and why Claude, Gemini and ChatGPT feel different
  • Hands-on: Token economics calculator, embedding space exploration, temperature experiments
  • Why AI Fails: hallucination, context overflow, prompt injection, sycophancy, and how to prevent each
  • Architecture Decision Matrix: when to use prompt engineering, RAG, fine-tuning, or agents
  • Hands-on: Hallucination forensics on your own domain; group architecture design challenge
  • From Prompt to Agent: system prompts, few-shot learning, chain of thought, tool use, building your first agent
  • 60-second MCP teaser (dedicated standalone workshop available separately)
  • Business Decision Framework: model comparison table (OpenAI, Claude, Gemini, Kimi K2.5, GLM 5, Qwens), MIT/Harvard productivity data, personal AI ROI worksheet
  • The OpenClaw disaster: 150K+ GitHub stars, 4,500+ exposed instances, plaintext credentials, zero authentication
  • The Lethal Trifecta: why personal AI assistants become security nightmares (private data + network access + untrusted input)
  • Hands-on: Building the secure foundation in four layers: isolated execution environment, authentication gateway, human-approval confirmation flows, network egress allowlist
  • Adding high-value capabilities: persistent memory, messenger integration, connecting to real-world data sources (specifics adjusted based on pre-workshop survey)
  • Testing your defenses: prompt injection, credential exfiltration, unauthorized network egress. Attack your own system. Fix what breaks.
  • What you take home: working secure AI assistant, security validation checklist, architecture decision matrix, personal AI ROI workshee
  • Bridge to Day 2: your assistant becomes a case study for Bogdan's cybersecurity sessions
  • Cozy space for networking and relaxing: Tea, kava, soft drinks, pizza, fire pit.

AI & LLM Platforms:

  • Claude API (Anthropic)
  • ChatGPT API (OpenAI)
  • Google Gemini API
  • OpenAI Tokenizer
  • Hugging Face Transformers

Development & Execution:

  • Docker (isolated execution environment)
  • Python 3.x
  • Node.js / TypeScript
  • VS Code / Cursor IDE
  • Git / GitHub

Security & Authentication:

  • OAuth 2.0 authentication libraries
  • JWT (JSON Web Tokens)
  • Environment variable management (.env)
  • Network firewall configuration tools
  • SSL/TLS certificates

Agent Framework & Tools:

  • LangChain or LlamaIndex (agent orchestration)
  • MCP (Model Context Protocol) - teaser
  • Custom prompt engineering templates
  • Function calling / tool use libraries

Testing & Validation:

  • Prompt injection test suites
  • Credential scanning tools
  • Network monitoring utilities
  • Security validation scripts

Integration & Capabilities:

  • Telegram Bot API / Discord API / Slack API (messenger integration)
  • Redis or SQLite (persistent memory)
  • REST API clients
  • Webhook handlers

Analysis & Worksheets:

  • Token economics calculator (custom tool)
  • Embedding visualization tools (t-SNE/UMAP)
  • Temperature testing sandbox
  • Personal AI ROI worksheet (Excel/Google Sheets template)