Mathias Biilmann Christensen

If your internal APIs live in tribal knowledge, good luck scaling. 🏗️ The Private API Network brings discovery and structure to internal work...before someone rebuilds what already exists. 🧭 Explore how searchable APIs save time and prevent duplication. https://lnkd.in/gBic8bAp

19

1 Comment

One of our customers from Tilt (formerly Empower) just shared how they wired the Augment CLI directly into their CI/CD pipeline for PR reviews. It’s a great example of what “agentic engineering” looks like in practice. Some lessons 👇 Their pipeline flow: - GitHub PR triggers Azure Pipeline - Pipeline hydrates a Docker image with PR code + metadata - Augment CLI runs review stages with access to MCP servers (Linear, LaunchDarkly, Notion, GitHub) 💡Lesson 1: Treat AI rules like code. They maintain “always apply” rules just for Augment. When something breaks, they triage it in Slack like a bug. Most teams fail by treating AI feedback as vibes. They treated it as engineering. 💡Lesson 2: Separate rules for humans vs. AI. - Humans get full docs in Notion (“why the rule exists”). - AI gets short enforceable rules (“always do X, never do Y”). The split makes both groups more effective. 💡Lesson 3: Hydrate context. With Notion + Linear + LaunchDarkly hooked up, Augment doesn’t just say “this violates a rule.” It explains why the rule exists, or checks tickets/flags directly. Feedback becomes coaching, not nagging. 💡Lesson 4: Modular prompts. They didn’t dump everything into one mega-prompt. Instead, the CLI runs in stages: - Code review - QA notes from Linear - Feature flag validation 💡Lesson 5: Centralize and version rules. - Rules live in the repo, next to the code they govern. - Same rules apply in IDE, CLI, and pipeline. No hidden logic. Developers see exactly what Augment enforces. The result: faster, higher-quality reviews that also teach developers the why. This is what agentic engineering feels like: less typing, more orchestration, agents enforcing consistency while explaining the reasoning. Big thanks to James Garrett for sharing this.

86

7 Comments

While everyone was talking about Opus 4.6, for me the true killer feature of the recent Claude Code updates is the agent teams. It's been something i've been trying to achieve with customization for a while. Custom agents, orchestration scripts, specific CLAUDE.md instructions to coordinate work... with some degree of success. But what Anthropic shipped natively is a WHOLE different level. What makes this stand out is the inter-agent communication. We're not talking about simple fan-out/fan-in where you spawn workers and collect results. These agents talk to each other. Peer-to-peer messaging, dependency-aware task graphs that auto-unblock, agents that self-claim work from a shared task list. The lead can even enter Delegate Mode where it does ZERO implementation, only coordination. The image below is from one of my setups. A team manager orchestrating a librarian agent, a PhD lead, and 5 research sub-tasks with blocking dependencies. The librarian unblocks the research tasks, the PhD lead aggregates everything. All coordinated autonomously. And with this a whole new world of orchestration just unveiled itself. Distributing work across agents is the "easy" part. You break down tasks, assign owners, define dependencies. The HARD part, and what MOST stands out now, is aggregation. How do you take the output of 5 parallel agents, each with their own context window, and synthesize it into something coherent? That's the new skill. Anthropic themselves used 16 parallel agents to build a 100,000 line Rust C compiler that compiles the Linux 6.9 kernel. No human actively coding. ~$20,000 in API costs over ~2,000 sessions. We went from pair-programming with AI to managing AI engineering teams. The skills that transfer are the ones from engineering management: task decomposition, context management, knowing when to intervene vs let the team self-organize. This is a new paradigm, and i think it opens up several possibilities we haven't fully explored yet. ref.: https://lnkd.in/dVCe344z

68

12 Comments

LLMs can now drive Playwright directly, generate tests, refactor them, and even monitor production apps... or can't they? Here’s how far we got with Playwright MCP and AI-powered workflows 👇 https://hubs.ly/Q03PgXyD0

12

Chunking and snippet extraction has been a huge focus lately - my latest blog dives into some of the work we've done on this to date, including support for a chunk rescorer in our semantic reranking retriever, as well as some useful ES|QL primitives to get more visibility into chunks and snippets. #elasticsearch #semanticreranking #snippets #chunks #chunking #esql https://lnkd.in/e6UD7iii

13

Tripadvisor's solution was originally built using Cassandra on-prem. But as their scale increased, so did the operational burden. After running a Proof of Concept with #ScyllaDB, the throughput was much better than Cassandra, and the operational burden was eliminated. We look at the technical specs here > https://ow.ly/5YMO50WkjP5

25

🚀 New on the Honeycomb blog: Unlock real-time visibility into your SaaS tools. With Webhookevent Receiver, you can now ingest events from tools like GitHub, Auth0, and Vercel—without writing custom code. This post walks you through the step-by-step process to get started. 👉 Read it here: https://lnkd.in/eihSUCNJ #Webhook #OpenTelemetry #Kubernetes

15

Indeed's counterintuitive API strategy: fund two teams to build competing GraphQL platforms, choose the winner based on real performance data. The result? OneGraph now processes 100 billion requests per month across 295 subgraphs, powering Indeed's global marketplace and external partner integrations. Mike Cohen, Technical Fellow at Indeed: "Like all great software companies, we built it twice." The lesson for engineering leaders: sometimes the fastest path to the right solution involves building competing approaches first. Swipe to see how Indeed turned duplication into discovery →

15

So which models produce HTML that has the least accessibility errors, and at what cost? Mapped it out here (bottom left is the best). This is helpful for understanding which models are most compliant for those with disabilities and do it efficiently. Qwen and Gemini 2.5 Flash lead the pack. Hat tip to Ben Ogilvie for pointing me to this: aimac.ai

55

2 Comments

In a new blog post, Marc Bowes looks at how updates work in Aurora DSQL, and what it means for scalable schema design. Read it here: https://lnkd.in/gmGEK8Dv A couple take-aways for application builders: - You mostly don't have to worry about hot read keys in DSQL, even in read-write transactions. DSQL can scale out per-key read throughput, and the design requires no co-ordination between transactions reading the same key. (Thank you, MVCC and Time Sync!) - Scalable applications do need to avoid high-contention write keys when doing updates. Writes need to be coordinated and ordered (to implement the I in ACID), which limits per-key throughput. - Most applications can avoid hot write keys with the right schema design. Most applications have already been written this way, because high-contention keys perform poorly on relational databases of all kinds.

240

11 Comments

Any best guesses on how long until Claude Code can: 1. Fire up the browser (without needing the MCP configuration) to evaluate changes 2. Fully leverage the Chrome Dev Tools 3. Use the Ruby debugger I’m finding Claude Code is creating fixes that would have been too tedious or painful. The gap right now is testing and debugging.

10

4 Comments

On behalf of the entire Percona product team for MongoDB, we're excited to announce a significant enhancement to Percona Server for MongoDB: File Copy-Based Initial Sync (FCBIS). It's designed to accelerate your large-scale database deployment with a more efficient method for initial data synchronization, reducing the time and resources required by the initial sync process. https://hubs.ly/Q03zL3qj0

14