Workshop proposal · The AI Capitol

Build Your
Hermes AI Agent
to be your
Capable TeammateFor Singapore Institute of Technology

A hands-on SIT session where participants leave with a working AI agent they own and control.

Lionel Sim & Benjamin Cheng

The AI Capitol

Who we are

Lionel Sim

Founder, The AI Capitol

Organises large-scale AI education programmes and agent communities in Singapore.

Founder of The AI Capitol and organiser of OpenClaw Singapore, 5,000+ members.
Former senior leader at Apple, TikTok and Tencent.
Education: Stanford, Cornell, Harvard, Wharton.
Honours: AdAge 40 Under 40; author of “The AI Agent Revolution.”
Connect: LinkedIn

Benjamin Cheng

Finance, Operations & AI Automation

Regional leader across APAC and UAE with background in group finance, systems implementation, audit and AI automation.

Chartered Accountant of Singapore, ISCA. ASEAN CPA. Certified Xero Advisor.
14 years of finance-related experience across group finance, systems implementation, audit and AI automation.
Has run several AI workshops and talks for professional and community audiences.
Education: Monash University, ISCA Singapore and CPA Australia.
Experience across Singapore, Hong Kong, China, Malaysia, Australia, New Zealand and UAE/Dubai.
Connect: LinkedIn

Overview

A working agent, not a slide deck

SIT students and staff from every discipline leave with a working personal AI agent they can use for study, projects, research or daily planning.

This beginner-friendly workshop combines a short explanation, live demonstration and guided hands-on practice. Participants do not need coding experience. By the end, each participant will have set up or accessed a working Hermes agent, run one practical workflow and scheduled one repeatable task using sample data only.

Concept

What is an AI agent

A chatbot answers one question. An agent does the work, remembers the context, and runs the task again on a schedule.

Chatbot

One question, one reply, no memory, no follow-through.

Agent

Long-running runtime: memory, tools, skills, channels and scheduled tasks.

Why Hermes Agent

It remembers and improves - persistent memory means the agent learns from prior sessions instead of starting cold every time. Hermes Agent self-improves and updates skills autonomously.
It self-learns repeatable workflows - participants give instructions once, then the agent schedules the task and improves from results.
It works across disciplines - the same foundation can support study, research, projects and future Business, Design, Finance, FoodTech, Healthcare or Engineering sessions.
Students and staff can build something useful - not just another prompt or chatbot conversation.

Impact

What it potentially brings to SIT

Agent literacy is becoming a core competency, not a novelty.

For students

A practical foundation for students from all disciplines including ICT, nursing, physiotherapy, hospitality, accounting, chemical engineering and foodtech.

For staff

A practical way for faculty and professional staff to explore agentic AI for teaching support, administration and research.

For SIT

A scalable introduction to agentic AI that can be aligned to SIT's approved devices, models, channels and data-handling requirements.

Proposal

The proposal

Build Your Hermes AI Agent to be your Capable Teammate - Foundation

Understand what an agent is and why it matters.
Install Hermes in a sandbox.
Connect a model and messaging channel.
Build and test a first workflow.
Leave with a running agent and adaptable templates.

Optional

Optional Track (For reference)

Build Your Hermes AI Agent to be your Capable Teammate - Intermediate

Configure SOUL.md and fallback models.
Set up a central gateway for local and cloud models.
Get started with local LLMs on your own hardware.
Understand agentic loops and Kanban.

Outcomes

Learning outcomes

Explain the difference between a chatbot and an agent.
Set up or access a working Hermes agent without needing to code.
Give the agent a clear role and instruction for a practical task.
Run one workflow for study, research or project planning.
Schedule one repeatable update or task.
Test for inaccurate or unsafe output and know what data not to use.

Use cases

What students can build

Foundation workflow	Student outcome
Study and deadline planner	Turn module dates and tasks into a realistic weekly plan.
Research with source links	Ask a question, receive linked sources and flag uncertainty.
Project planning assistant	Turn notes into tasks, owners, deadlines and open questions.
Daily industry briefing	Follow a field such as healthcare, finance, foodtech or engineering.
Future domain pathways	Extend the generic foundation into approved Finance, FoodTech, Healthcare or Engineering sessions.

Format and logistics

How we deliver

Format

70% build, 30% concept blocks, live demo, guided hands-on, pair support.

Duration

Recommended: one Foundation session, followed by optional homework review.

IT environment

Participants use their own Windows or macOS laptops (minimum macOS 14 on Apple Silicon M-series chips). Docker Desktop installation is required to run Hermes securely in a sandbox. This is the recommended and safest path: Hermes runs inside an isolated container, not directly on the host.

Security

Foundation uses publicly available synthetic data only.

Timeline

Build Your Hermes AI Agent to be your Capable Teammate - Foundation session: five hours

Time	Segment	Outcome
00:00 - 00:30	What agents do	Understand the difference between a chatbot and an agent.
00:30 - 01:30	Guided setup	Access Hermes in Docker or an approved fallback environment.
01:30 - 02:00	First conversation	Give the agent a role, instruction and safe sample task.
02:00 - 03:00	Build one workflow	Create a study, research or project-planning workflow.
03:00 - 04:00	Schedule and test	Schedule one repeatable task and check the result.
04:00 - 04:45	Use cases and safety	Choose a next workflow and understand data boundaries.
04:45 - 05:00	Recap and homework	Leave with a working agent and one guided challenge.

Proof points

Proven in real sessions

Two five-hour hands-on builds on 27 June and 18 July 2026. Participants left with a running agent.

FEEDBACK · 18 JUL 2026

“I took the leap and built my first Hermes AI Agent... The learnings were not only about the technical aspects of Hermes setup, but also best practices to save token costs, secure your implementation, and pitfalls to avoid.”

— 彭惠文 Evelyn Phang, Partnerships & Ecosystem Leader

FEEDBACK · 27 JUN 2026

“Yesterday was an eye opener on just how far behind I am. The way it was structured with clear learning objectives helps a lot with keeping up and building confidence on how to use the tools.”

— Aaron, workshop attendee

Pricing

Pricing for Build Your Hermes AI Agent to be your Capable Teammate - Foundation

Cohort30 pax

Special promoS$9,990 / session

Original priceS$15,995 / session

Inclusive	Not included
Facilitation, slides, setup scripts, sample workflows, post-session references	Participant laptops, Docker Desktop, model API keys

Payment terms: Payment to be made in two instalments. First 50% professional fees to be paid upfront upon confirmation and balance 50% professional fees to be paid within 30 days after course completion.

Decision points

Next steps with SIT

Agree the Foundation Session date.

Thank you!

Build YourHermes AI Agentto be yourCapable TeammateFor Singapore Institute of Technology