Spec Coding & Agentic Engineering with Claude Code - Hands-on Workshop
What Is This Workshop?
This is a 5-hour, hands-on workshop on Spec Coding and Agentic Engineering using Claude Code — held in-person in Singapore's CBD on 24 June 2025, 1pm–6pm.
If you've been experimenting with AI coding tools, vibe coding, or copy-pasting prompts into ChatGPT or Claude — and wondering why your projects break the moment you try to scale — this workshop is built for you.
You'll learn the exact methodology that separates developers who build reliable, AI-powered software from those stuck in an endless loop of fixing hallucinated code.
Seats are strictly limited to 20 participants.
What Is Spec Coding?
Spec Coding is a structured, specification-driven approach to AI-assisted software development. Instead of casually prompting an AI and hoping for the best (vibe coding), you write clear functional blueprints that guide Claude Code — or any AI coding agent — to produce consistent, scalable, and maintainable output.
Spec Coding is the foundation of Agentic Engineering: directing teams of automated AI agents to write, test, and fix code as a coordinated pipeline — not as a one-shot prompt.
This workshop teaches you both.
What You'll Learn (Workshop Agenda)
1. Moving Beyond Random "Vibe Coding" Understand why copy-pasting prompts fails at scale and what separates vibe coding from a repeatable engineering process. You'll leave with a mental model that changes how you interact with AI development tools forever.
2. Writing Clear App Blueprints Master the art of writing functional specifications — structured documents that tell Claude Code exactly what to build, how to handle edge cases, and what the expected outputs are. This is the core skill that unlocks reliable AI-generated code.
3. Building Your Live Portfolio Website Apply spec coding in real time to deploy a personal portfolio website. You will write the spec, run it through Claude Code, iterate, and publish — all within the session.
4. Building a Smart Mini-CRM App Level up by tackling complex application state and data persistence. You'll build a functional mini-CRM app, confronting the exact challenges that trip up most AI-assisted development projects — and solving them systematically.
5. Working with AI Developer Teams Go beyond single-agent prompting. Learn how to direct multiple automated AI agents to write different parts of your codebase in parallel — the way professional agentic engineers work.
6. Running Test-Driven AI Pipelines Learn agentic engineering techniques to automatically test, identify bugs, and trigger fix cycles — without manual intervention. This is how you ship software that actually works.
Who Should Attend?
This workshop is designed for:
Startup founders who want to ship products faster without a large engineering team
Seasoned developers looking to multiply their output using AI coding agents
Software engineering leaders exploring agentic workflows and AI-native development
Product managers who want to build and prototype without being blocked by developers
Students and early-career technologists who want to be AI-native engineers from day one
No advanced coding background required — but participants should come prepared to build. All participants are expected to have an active Claude Code subscription.
Your Instructor: Santosh Rout
Santosh Rout is the founder of navan.ai and Saaragh, with over 25 years of experience in the IT and AI industry.
He is the creator of sam.navan.ai — a leading Agentic Spec Coding platform — and has helped countless founders and engineering teams bridge the gap between complex product ideas and working, scalable software.
Santosh's expertise sits at the intersection of agentic workflows, specification-driven development, and production AI systems. He has built multiple successful AI products and brings hard-won, practitioner-level knowledge to every session.
What Past Participants Have Built
After attending navan.ai's spec coding and agentic engineering workshops, participants have gone on to build:
A Peer-to-Peer Delivery Platform — allowing travellers to list journeys and get paid to carry parcels along their route
An EdTech Platform for Extra-Curricular Tracking — helping students record and measure the impact of volunteering and extracurricular activities
An AI-Powered CCTV Planning Tool — where users upload floor plans and the system automatically suggests optimal camera placements and generates real-time quotes
These are not toy projects. These are production-grade apps built by non-engineers using spec coding.
What a Past Participant Said
"Thank you for taking the time to conduct the session on Agentic Spec Coding. The discussion gave me a much better understanding of the concepts and practical approach involved. I especially found the insights very helpful. I look forward to applying what I learned and continuing to improve my skills in this area." — Anitha P, Virtual Participant.
View snapshots of our previous workshops, bootcamps, and hackathons: https://asca.navan.ai/gallery/
Event Details
Date: Wednesday, 24 June 2025
Time: 1:00 PM – 6:00 PM SGT
Format: In-person, hands-on workshop
Location: CBD / Central Business District, Singapore (exact venue will be shared with registered participants in a few days)
Seats: Limited to 20 participants
Prerequisite: Active Claude Code subscription
Fee: SGD 200
Organiser: Navan.ai
Frequently Asked Questions
Do I need to be a developer to attend? Not necessarily. The workshop is designed for founders, PMs, and builders — not just developers. A basic comfort with technology helps, but the methodology is accessible to anyone willing to engage actively.
What do I need to bring? Your laptop and an active Claude Code subscription. Come ready to build.
Why is this better than just watching YouTube tutorials on Claude Code? Tutorials show you what's possible. This workshop gives you a repeatable system — spec coding methodology — that you can apply to any project, of any complexity, without breaking things the moment requirements change.
What makes spec coding different from vibe coding? Vibe coding is ad-hoc: you describe what you want loosely and hope the AI figures it out. Spec coding is structured: you write a functional blueprint that the AI executes deterministically. The output is modular, maintainable, and won't collapse when you need to iterate.
Is this workshop specific to Claude Code? The hands-on work uses Claude Code, and participants will need a Claude Code subscription. However, the spec coding methodology taught is platform-agnostic — it applies to any AI coding agent.
Organised by navan.ai — the agentic spec coding platform powering the next generation of AI-native builders.
Got questions? Reach us on connect@navan.ai or call us on (+65) 8608 9171.