[Operations Edition] Building an Investment Research Bot for Fund Managers — The Art of Running a Bot Reliably for the Long Haul

We build seven practical investing bots, from U.S. market close briefings and earnings alerts to target-price radar and interest-rate triggers. We make them robust with shared components, connect them into a single system, and run them automatically every day.

I’ll build an investment research bot that automatically delivers a market briefing, alerts on sharp moves in your watchlist, and keyword-based news summaries via Telegram every morning at 6 a.m. Based on a design actually used by a fund manager overseeing a 100-billion-won fund, I’ll set up free infrastructure that runs 24/7 without a server and build three bots from scratch.

There's so much talk about AX these days. At Selfish Club, we started asking ourselves, "How 'agentic' can our way of working actually become?" So, for two months, we just dove right in. From how we drew the roadmap and what our non-developer crew members built with Claude, to how we transformed the process of creating product detail pages—which we used to build from scratch every time—into a system. Instead of a polished conclusion, we're sharing the raw, chaotic process of our trial and error.

You can now watch the online version of the offline meetup where Jae-kyu Lee—developer of the open-source Agent OS 'Ouroboros,' which quadrupled its GitHub stars in just three months—shares design principles gained from directly integrating and dismantling eight coding agent runtimes!

We are revealing the entire process—without filters—of how non-developers built a Slack bot from scratch, including all the trial and error along the way.

I will show you "where to start" through a case study of building a service with non-developers in just one day, and explain "how to manage and collaborate" using the example of designing the GitHub structure for the Selfish Membership team.

The lecture explains why the harness is more important than the model through the lens of Claude Code architecture. It also covers how to design a Claude Code work system around six structures: Context Harness, Permission Harness, Verification Harness, Tool Harness, Subagent & Worktree Harness, and Debugging Harness. It then connects these concepts to CLAUDE.md, session management, controlling large codebases, parallel worktree operations, error log analysis, and tracking environment variable and configuration scope issues, helping you establish practical standards for continuing to use Claude Code in real-world backend and DevOps work.

Have you introduced AI but still find yourself losing time to repetitive tasks? The problem isn't the AI; it's the 'system.' This course guides you through the process of building a team of six AI assistants—including an executive assistant, content assistant, marketing assistant, CRM assistant, and personal productivity assistant—all using Notion AI. Based on my experience automating hundreds of repetitive tasks in real-world business settings, we will build actual workflows from start to finish, covering everything from organizing notes and automatically generating meeting minutes to blog writing, CRM construction, and automated KakaoTalk notifications. After completing this course, you will have a system where your own AI assistant team works automatically, powered solely by a Notion Business plan.

From meeting preparation to writing minutes, AI does it all! The first in our series on using AI Agents for company work!

We invite you to become a ‘Top-Tier Enterprise AI Systems Architect’ Do not remain a micro-level technician (Prompt Writer) who merely assembles libraries created by others or polishes prompt sentences. From mathematical first principles to runtime control-loop kernels and autonomous recovery infrastructure — through 40 years of practical engineering insight that has driven industrial operations in South Korea, I will firmly guide you to become a top-level ‘Enterprise AI Systems Architect’ who protects your company’s business and digital sovereignty. Now, begin the great journey of transforming uncontrollable AI into a perfect industrial core. “Why does an AI that ran perfectly in a demo start spouting strange nonsense or bring the pipeline to a halt as soon as it is deployed to production?” “Why do AI projects that so many companies launch with great enthusiasm ultimately fail to escape the ‘PoC (Proof of Concept) swamp’ and get abandoned?” These are the most desperate questions that countless developers and systems architects I have met in the field have asked me. Hello. I am Hongjip Kim, who spent more than 40 years in the powertrain R&D division of the South Korean automotive industry (including 17 years as an executive in Hyundai Motor Group R&D), led the export of engine and drivetrain technologies worth 130 billion won, and commanded operations as the CEO/CTO of a listed mid-sized manufacturing company. From the perspective of an engineer who has spent decades in mechanical engineering and precision-control environments, a ‘risk of fatal uncertainty’ lurks behind today’s frenzy to adopt generative AI. In mission-critical industrial environments where even a mere 0.1% error or mistake can directly result in the destruction of physical equipment worth hundreds of millions of won and an enterprise-wide shutdown, integrating a probabilistic chatbot that simply relies on probability to generate plausible text directly into a service is like carrying a ticking time bomb into a fire. This course was designed to impart the “practical engineering methodology for training uncontrollable probabilistic AI into the most robust industrial control core, one that allows not even 0.000% of syntax errors.” 1. The 3 Major Fatal Bottlenecks You Face in the Field (Problem) Many engineers attempt to challenge production environments using an approach limited to one-off prompt tuning (Vibe Prompting), only to become frustrated when they encounter the following 3 structural defects: Finite Context and ‘Lost in the Middle’: As conversations or multistep executions grow longer, critical system instructions are pushed aside and the context window is exhausted, causing the system to lose its memory and amplifying hallucinations. Nondeterministic Outputs and Cascading Pipeline Collapse: Tiny probabilistic fluctuations, such as one missing closing bracket in JSON or an incorrect type, can paralyze legacy ERP/CRM API calls and destroy data integrity. Untrusted Code Execution and Infrastructure Security Bankruptcy: When scripts autonomously generated by an agent run in a non-isolated host OS environment, they create risks such as kernel escapes and sensitive-data leakage. As such, AI malfunctions in industrial environments are not merely quantitative accuracy problems; they become a massive financial disaster in the form of asymmetric business loss (Asymmetric Business Loss). 2. ‘How’ to Solve These Problems Perfectly (Solution) This course differs fundamentally from commercial courses that teach only how to use fragmented off-the-shelf tools. By combining 40 years of hands-on field expertise with computer systems engineering, it delivers 3 hardcore architectures for vertically integrated, system-level control of probabilistic AI: ① Move Beyond Prompting to ‘LLM-OS & Context Orchestration’ Following Andrej Karpathy’s paradigm, redefine the LLM not as a conversational respondent but as the central computational core (CPU) of a computer. Systematically organize the context window as ultra-low-latency volatile RAM, the internal vector DB as Disk, and external APIs as an I/O bus. Through MemGPT hierarchical virtual-memory paging, preserve critical state without saturation even during infinite conversations, while scaling a dynamic agent network through MCP-based Liquid Swarm MAS, minimizing token costs. ② ‘Deterministic Harness Engineering’ That Constrains Probabilistic Hallucinations to 0.000% Leave the language model’s creative flexibility open only during the Reasoning phase, while firmly confining the physical output layer that interacts with external systems using mathematical Boolean algebra and compilers. Apply EBNF / XGrammar-based CUDA logit masking to forcibly block the generation probability of invalid tokens at P = 0.000%. Form a physical security isolation ring so that code generated by agents runs only once inside a Firecracker MicroVM micro-virtualization sandbox with 5ms snapshot restoration capability. ③ ‘High-Reliability MLOps·LLMOps Autonomous Recovery Infrastructure’ Without Wasting GPU Resources Build vLLM PagedAttention and S-LoRA hot-swapping to suppress idle VRAM waste to below 4%. When a multi-agent API fails, restore data consistency 100% through SagaLLM reverse compensation transactions. Through an OpenTelemetry/Langfuse distributed observability network, defend system availability by executing an unmanned automatic canary rollback whenever a decline in the RAGAS faithfulness score (Data Drift) is detected. 3. Practical Value That Drives Spiral Growth (Pedagogy & ROI) This course is systematically designed according to the 5 levels of Bloom's Taxonomy. Starting from simple theoretical recall (Remember), learners directly code and validate a 5-stage debugging pipeline (analyzing deceptive baseline trap source code ➔ deconstructing adversarial engineering ➔ establishing a Robust integrity pipeline ➔ calculating the TEI ROI formula ➔ SOTIF/ISO 42001 governance). Going beyond the intuitive satisfaction of simply “working well,” I will apply the Forrester TEI (Total Economic Impact) financial model to visualize, from an executive perspective, how much the AI system you build reduces your company’s rework cost (Rework Cost), avoids failure risks worth hundreds of millions of won, and proves an ROI of more than 200%.

How much can you entrust to an AI Agent? IBM engineers answer with evidence they’ve put to the test, covering everything from LLM fundamentals to vendor-specific design, security, and evaluation—and help you create a Skill to take over your work.

Instead of writing longer prompts, I create rule files, logs, permissions, hooks, and validation routines one by one whenever something goes wrong in a real project. With a news alert bot that I actually built and deployed to a server in five days, 211 screenshots and 87 instructions exactly as originally written.

AI handles the first draft; I handle the judgment and polish. Learn the practical skills and terminology that junior planners use from day one, and work with Claude to create a complete set of deliverables for a study café reservation service: a project proposal, menu structure, screen designs, and functional specifications.

From AWS EC2 backend operations to automatic error fixing, we are revealing the entire development process of Sidabari (https://github.com/cx8537/sidabari), a Claude Code-based automation tool. In this course, we provide all the actual prompts and conversation history used as learning materials, allowing you to experience the entire prompt input process involved in its creation.

Latest version as of October 2026. This is a 299-lesson course created by YouTube’s “AI Adjunct Professor Lee Jong-beom” (54,000 subscribers). It guides you step by step, starting with how to open a browser tab and progressing to creating an AI employee team with Claude Code. It includes approximately 290 practice worksheets and 1,000 credits.

Do you still code by hand these days? We’re in an era when even vibe coding feels like a chore! Become an AI automation expert with n8n—no coding required. A guide to designing automation workflows you can put to work right away Practical, hands-on automation strategies to maximize efficiency. Build intuitive workflows with n8n and experience a revolution in the way you work.

The era of simply receiving code "suggestions" is over. I will clearly and simply introduce you to Google Antigravity, a next-generation agent development platform where AI agents directly control the editor, terminal, and browser to plan, execute, and verify complex software tasks.

From tokens and context to RAG, agents, and guardrails, we examine how AI services work from a security perspective. We compare the concepts with the provided backend code and organize the areas to inspect across input, retrieval, and tool calls into a single attack-surface map.

Can you make a proposal look convincing with AI, but lack the confidence to bring it into the meeting room? A global content strategist with 17 years of experience—who built practical planning expertise at major advertising agencies and global e-commerce companies, and had that expertise validated on a global education platform—will teach you the principles of planning that work in practice and AI workflows, from defining the problem to completing the slides.

Test the agent’s authorization boundaries for tool calls and RAG search and document access. Using the actual model and training mock tools and documents, observe the approval gates and search permission filters, then complete the Mission 5·6 diagnostic worksheet.