Boundaries That Hold & Context That Scales @ FrontValue

FrontValue is hosting the next Dutch Angular Meetup. Join us to hear great talks, connect with like-minded people, and share dinner and drinks!

💡 We kick off the evening with Roberto Heckers.
You now him from his insightful and very interesting (blog) posts on social media and talks at Angular events in the EU.

Everything I need to know about monorepos I learned in kindergarten
A story about sharing and boundaries

Every kindergarten class has two kids. One hoards the blocks and won't let anyone near them. The other gives away all his toys and goes home with nothing. Monorepos have both, and neither of them is having a good time.

Sharing code is why we build monorepos in the first place, but shared code without boundaries turns into a shared folder that owns everything and can't be changed by anyone.

In this talk we'll walk through an Nx workspace structure that makes boundaries obvious before you enforce them, then set up those boundaries properly with tags and dependency constraints. From there, we'll look at the sharing decisions that actually come up day to day, and at what separates code that's genuinely shared from code that just ended up in the same folder.

💡 Next on stage is our own Frank Merema.
As a seasoned frontend consultant, he has worked in several complex organizations and projects, mainly with Angular. He also serves on the FrontMania conference program committee.

Beyond Context Windows
How we cut LLM token usage by 46% with a Neo4j Code Graph

Large codebases don't fit in an LLM's context window — and when you can't fit the whole repo in, the AI guesses. Wrong file, wrong dependency, broken refactor.

In this talk, I'll walk through how we replaced "stuff everything into the prompt" with a Neo4j knowledge graph of our entire codebase — files, functions, classes, imports, and call graphs, all linked by their real relationships. Instead of feeding the AI everything and hoping, we let it write Cypher queries to pull exactly the slice of the codebase it actually needs.

I'll cover the architecture end to end, including the MCP layer we built in front of the raw graph to combine semantic search with graph queries, control user input, and produce answers that are more stable, smarter, and cheaper — plus the lessons learned getting there.

The results from our early runs: 46% fewer tokens per query, 79% faster response times, and answer accuracy up from 15% to 59%.

If you're fighting context window limits or token costs with LLMs on large codebases, you'll walk away with a concrete architecture to try yourself.

📍 Location
We welcome you to our office in Utrecht: Studio N in the Werkspoorkathedraal. The address is Tractieweg 41. We have plenty of free parking space available. We will have dinner ready for you at 18.00!

🕔 Agenda
18:00 Walk in & food
18:30 First talk
19:30 Break
20:00 Second talk
21:00 Drinks & Bites

Both talks are in English. RSVP now for an evening of learning and fun.
Don't miss out on this amazing opportunity to learn and connect with our community of frontend developers!