Celebrating one decade of Algoritma!

Next cohort starts 11 January 2027

Learn MCP andAI agentsin 2 weeks

A live bootcamp for analysts, marketers, operations staff and developers. You start with how assistants like Claude and ChatGPT work, connect them to real tools through the Model Context Protocol, and finish by building and deploying your own MCP server. AI subscriptions and a ready-made workspace are included.

US$290 or IDR 5,298,590 per seat, with AI subscriptions and tooling included.
The next cohort runs from 11 to 24 January 2027.

2

weeks bootcamp

8

live evening lessons

5

MCP servers

6

projects for your portfolio

your bootcamp workspace5 MCP servers
Claudeyou, day 8
Sectors
Viably
Filesystem
GitHub
Your server
readfilesystem.read_file

What you work with

ClaudeClaude CodeChatGPTCursorCodexModel Context ProtocolSectorsViably
SlackGitHubGoogle CalendarStripeGoogle Search ConsoleInstagramFacebookThreadsViably OCRWebhook RelayEmail
MCP InspectorTypeScript SDKPython SDKStreamable HTTPOAuthNode.jsPrompt librariesAgent skillsCLAUDE.md
Who it's for

Built for professionals in any role

Most of the cohort has never built an AI integration before. The course starts from zero and gets advanced quickly, with instructors in the room for every lab.

Changing careers

You want to move into AI-related work and need hands-on experience you can show. You leave with six projects, a deployed MCP server and the vocabulary to talk about agents in an interview.

Adding AI to your current job

You work in sales, marketing, finance or operations and want AI to take over the routine parts of your week: the inbox, the pipeline, payment reports, SEO checks and social reporting. You can use what you learn in week one at work the next Monday.

Developers new to agents

You can write code but haven't built with MCP or Claude Code yet. Week two takes you through server design, deployment and security at a working level.

Live lessons, Monday to Thursday

6:30 to 9:00 PM (GMT+7), after work. Eight lessons over the two weeks.

Practice day on Friday

No lesson. The cohort meets to work through the week's assignments together.

Self-paced at the weekend

Short lessons and exercises in your workspace, whenever suits you.

Online, in a small cohort

Instructors teach live and help you in the labs.
The two weeks

From your first prompt to your own MCP server

Each week has four live evening lessons, a practice day on Friday and self-paced work at the weekend. In week one you connect Claude to Viably and run your CRM, sales pipeline and Stripe payment reports through it. Week two moves into Claude Code for SEO fixes and a social media team, and ends with your own MCP server.

  1. 1

    Week 1, Mon, liveBasics

    How AI assistants work, and your first connection

    We start with what happens when you send a message: tokens, context windows, models and why answers vary. Then you connect Claude to your own Viably workspace over MCP and choose what it may read and change.

    • Tokens, context and model choice
    • Prompts with roles, examples and formats
    • Adding Viably as a Claude connector

    You build: Claude connected to your own Viably workspace

    Claude, a new chat
    What is a context window, in one sentence?

    context used1,200 tokens
  2. 2

    Week 1, Tue, liveBasics

    Run your CRM from a chat

    Claude works through the people waiting on a reply. It claims contacts, tags them, writes notes and drafts replies for a person to send, and you approve each change. You also learn why text from customers is treated as data and never as instructions.

    • Triaging the inbox with triage_inbox
    • Claiming, tagging and noting contacts
    • Approvals, scopes and prompt injection

    You build: An inbox triaged to zero, with drafted replies

    Claude, Viably connected
  3. 3

    Week 1, Wed, liveIntermediate

    Sales pipelines and deal tracking

    You set up a pipeline and let Claude keep it honest. It opens deals for new leads, moves them between stages, records why a deal was lost, and runs a weekly review that finds stalled deals and missed follow-ups.

    • Stages, values and win rates
    • Opening and moving deals with Claude
    • A weekly review of stalled deals

    You build: A Monday pipeline review that flags stalled deals

    Claude, sales pipeline
  4. 4

    Week 1, Thu, liveIntermediate

    Payment reports from Stripe

    Stripe sends every charge, invoice and subscription change to Viably, matched to the customer who paid. Claude reads those events and turns them into a monthly report with charts, then files a recovery task for every failed charge.

    • Stripe events in the Viably ledger
    • Revenue by customer and by source
    • Recovery tasks for failed charges

    You build: A monthly payments report, with recovery tasks

    Claude, Stripe events
  5. Week 1, Fri to SunPractice

    Practice day and a self-paced weekend

    Friday has no lesson. The cohort meets online to work through the week's assignments together, and you help each other get unstuck. Over the weekend you go through short self-paced lessons and exercises in your workspace.

    • Group assignments on Friday
    • Short self-paced lessons
    • Exercises in your own workspace

    You finish: Week one's assignments, finished with your cohort

    cohort workspace
  6. 5

    Week 2, Mon, liveIntermediate

    Troubleshoot SEO with Claude Code

    Week two opens in the terminal, which we teach from scratch. Claude Code reads your Search Console findings through Viably: pages losing clicks, queries that rank well but earn few clicks, and pages dropping out of the index. It fixes what it can in the site's files and files the rest as issues.

    • Terminal basics for non-developers
    • Search Console findings through Viably
    • Reviewing each fix before you accept it

    You build: SEO fixes applied, and the rest filed as issues

    terminal
  7. 6

    Week 2, Tue, liveAdvanced

    A social media team that works from data

    You build a small team of Claude Code agents: an analyst, a strategist and a writer. The analyst reads your Instagram, Threads and Facebook numbers through Viably and compares formats on the average post. The others plan next week's posts from what it finds.

    • Reach, saves and formats per platform
    • Subagents with one job each
    • A content plan backed by numbers

    You build: A three-agent team that plans next week's posts

    terminal
  8. 7

    Week 2, Wed, liveAdvanced

    Build and deploy your own MCP server

    You write an MCP server for a spreadsheet, API or tool Viably doesn't cover, test it with the MCP Inspector and put it online with a sign-in. Then you run one task that uses your server, Viably and Sectors together.

    • Designing tools a model uses well
    • Testing with the MCP Inspector
    • Deploying with OAuth and chaining servers

    You build: Your own server, live next to Viably and Sectors

    editor
  9. 8

    Week 2, Thu, liveAdvanced

    Demo day

    Each participant shows the workflow they built to the cohort and our instructors, and gets feedback. You finish with six projects in your portfolio and a certificate from Algoritma.

    • Live demos to the cohort
    • Instructor feedback
    • Next steps for your own work

    You build: Six portfolio projects and an Algoritma certificate

    demo day
Everything included

Your environment is ready before the first lesson

We provision every participant's tools, accounts and AI subscriptions ahead of day one. You don't install software or pay for AI subscriptions yourself.

provision --cohort next

Setting up your bootcamp environment

assign AI plans ............ claude, chatgpt

AI subscriptions

Paid Claude and ChatGPT plans for the full two weeks, so you never hit a free tier limit mid-lesson.

Cloud workspace

A coding environment in your browser with Node, Python and Claude Code already installed. Nothing to set up on your laptop.

Sectors API key

Access to the Sectors MCP server and API, which you combine with your own server in week two.

Viably workspace

Your own Viably workspace filled with sample contacts, deals, Stripe payments, Search Console findings and social accounts, so agents practise on realistic data.

Starter repository

A GitHub repository with the MCP server template you build on in week two.

Deploy target

Somewhere to put your MCP server online on day 7, with a URL you can keep sharing after the bootcamp.
Bootcamp instructors

Our instructors build AI systems at Supertype

Samuel built Sectors and Viably, the two MCP servers you use in class. Andreas and Lukas build AI systems for Supertype's clients. They teach live in the evenings and help you in the labs.

Samuel Chan

Samuel Chan

Founder of Algoritma, lead instructor

I am a 3-time software entrepreneur (currently Sectors, Algoritma and Supertype, with 1 successful exit). Between 2013 and 2017, I built and sold a company in chatbots, natural language AI and marketing automation while living between Beijing, Tokyo and my home town Singapore. I also contribute on GitHub, with many open source tools, libraries, and open courses on Large Language Models and Computer Vision.

  • Built software companies over two decades (Audience Kit, HyperGrowth (sold), Growth Bot (sold), Sectors Financial Data Suite, Viably) while leading engineering teams in Singapore, Beijing, Tokyo and Jakarta.
  • First in Southeast Asia to receive certification in Microsoft's Data Science professional program (2014) and in MongoDB (2014). Also holds Graph Analytics for Big Data from UC San Diego and Machine Learning by Stanford (2016), and was the first recipient in Asia of RStudio's certification.
  • Accredited instructor (NTUC Singapore, IBF Singapore) and guest lecturer at the University of Indonesia, Institute of Technology Bandung, Ngee Ann Polytechnic and Singapore Institute of Technology.
  • In-house consultant to SEGA-backed goGame, publisher of Disney Crossy Road and Big Hero 6, and to the Adaro (IDX: ADRO) group of companies, with past roles at DNP Group (TYO:7912), Gumi (TYO:3903) and DianDian FunPlus (600634.SH) in Beijing and Tokyo between 2011 and 2016.
  • Contributor to publications including The Logistics Institute at NUS (National University of Singapore).

17,000+

Students taught AI development since 2020

280,000+

YouTube views on his LLM and generative AI series

939 stars

On his LLM education series on GitHub (onlyphantom/llm-python)

2.2 million

People reached through his StackOverflow answers

40+

Open source scripts and LLM tools on GitHub

18+ years

Of software engineering, enterprise analytics and AI engineering

Andreas Christianto

Andreas Christianto

Knowledge Expert, RAG Systems at Supertype

Master's in Computer Science from Binus, with previous stints at Infracom and Toyota Astra Motor (DI/DX Big Data). Andreas is Supertype's in-house technical content writer. He develops technical documentation, knowledge base articles and instructional sessions for Supertype's RAG systems and AI solutions.

Lukas Wiku

Lukas Wiku

AI and Agricultural Intelligence at Supertype

Lukas has a degree in Agricultural and Biosystem Engineering and a Master's degree in Biosystem Engineering from Kangwon National University in South Korea. His skills span IoT, agricultural machinery and deep learning, with deep specializations in computer vision and aerial imagery.

Real servers

Two production MCP servers you use in class

You practise on servers that real teams depend on, so you can use the same skills at work.

Sectors

Financial data for Indonesian and Singaporean markets over a hosted MCP server with more than 40 tools. You use it in week two, next to the server you build. sectors.app

Viably

A work operating system with a CRM, deal pipelines, Stripe payments, Search Console findings and social analytics behind one MCP server. You use it from the first evening. viably.app
Next cohort

Join the January 2027 cohort

The next cohort runs from 11 to 24 January 2027, with live lessons from 6:30 to 9:00 PM (GMT+7), Monday to Thursday. A seat costs US$290 or IDR 5,298,590. We cap the cohort at 30 people so instructors can help everyone in the labs. You pay through Supertype, Algoritma's parent company, by card, QRIS or bank transfer.

Questions

MCP bootcamp FAQ

No. Week one uses Claude through its normal app. Week two starts with the terminal, which we teach from scratch, and later that week you build an MCP server with an instructor and Claude Code helping you. Developers move faster, and everyone finishes the same projects.

MCP, the Model Context Protocol, is the open standard AI apps like Claude, ChatGPT, Cursor and Codex use to connect to outside tools and data. Knowing how to connect, use and build MCP servers lets you turn an AI assistant into something that works with your actual files, data and apps.

The bootcamp costs US$290 or IDR 5,298,590. That covers paid Claude and ChatGPT plans for the two weeks, a cloud workspace with all the tools installed, a Sectors API key, your own Viably workspace with sample data, a starter repository, and somewhere to deploy your MCP server. You only need a laptop with a web browser.

Each lesson ends with the lab files in your workspace, so you can catch up over the weekend or on practice day and bring questions to the next lesson. Talk to us before you enrol if you know you'll miss more than one evening.

Yes. Participants who attend the sessions and present on demo day receive a certificate of completion from Algoritma.

Yes. Our Enterprise MCP Training runs privately for one company, with tracks for leaders, business teams and engineers and labs on your own systems.

Stay updated with Algoritma's latest news

We do not resell your data and respect your privacy.

Algoritma

Email us at [email protected] for enquiries on enterprise training, custom workshops, and AI initiatives for your team.

Enterprise Solutions

Algoritma | AI & Data Science Education

Programs & Company

Supertype Group of Companies

Supertype-Incubated Products

Information