How AI is changing software building at Port.io

Big moment at Port.io. Me and Yuki are proud to share the Agentic Engineering Platform (AEP) — built to help teams stay in control as AI becomes part of everyday engineering. The next era of software is here, and it’s built on trust, clarity, and collaboration between humans and AI. 👉 https://lnkd.in/dbjCtPy5

Thrilled to launch our Engineering blog, where Finch engineers will account first hand what they're building. The first one is written by none other than Moyez Mansoor, our Founding Engineer and Naya loyalist. Our first post dives into how we replaced our Celery/Django task queue with Temporal, beefing up observability 👀 and reliability 🗿 of the workflows that orchestrate our AI agents. Link in comments.

Behind every performant LLM is a hidden architecture of embeddings, namely vector spaces that decide how well your model understands context, language, and domain drift. At Innosquares, we’ve learned to classify embeddings by drift, decay, and feedback adaptability. This lets us optimise models for multilingual queries, dynamic documents, and domain-specific reasoning. It’s not the prompt. It’s the embedding map that makes or breaks generative quality. #VectorDatabases #Embeddings #LLMEngineering #RAG #AIInfrastructure

I’m more convinced than ever that the hardest part of modern engineering isn’t the tech; It’s the discipline. Anyone can bolt an LLM onto a product. Fewer teams take the time to design clean interfaces, evaluate behavior, and treat AI like a first-class part of their architecture rather than a “smart plugin.” Craft still matters. Especially now. 🛠️ #ModernSoftwareEngineering #AIEngineering #CleanCode #PlatformEngineering #AIGovernance #Observability #DeveloperExperience #EngineeringLeadership #AITrends

For 50 years, Stradling has been at the forefront of Orange County’s emergence as a leading hub for technology and innovation. From partnering with early hardware pioneers to counseling today’s AI, software, and life sciences leaders, the firm has helped shape every phase of the region’s tech evolution. Read the full article to learn how Stradling helped build Orange County’s technology legacy: https://lnkd.in/geMUGUiu

“We do not ride on the railroad; it rides upon us.” — Henry David Thoreau Echoed in the recent Autonomy vs. Empire essay by Harry Law, it is a powerful reminder that the systems we build do not just serve us. They start to define us. Every layer of infrastructure—railroads, algorithms, platforms—reshapes the way we move, think, and decide. We design them with purpose, but over time they begin to structure that very purpose. AI is today’s railroad. We believe we are steering it, but it is quietly rerouting how we work, create, and even understand autonomy. The question is not whether progress will continue. It is whether we will remain intentional about where the tracks lead. But like the railroad, the progress and impact that we are only beginning to see come to fruition. https://lnkd.in/eqhmSPHe

Kimi K2 Thinking is setting a new standard for open-source AI reasoning. Built on a trillion-parameter Mixture-of-Experts architecture, it supports ultra-long context (256K tokens) and can execute 200–300 sequential tool-calls—searching, coding, planning—for sustained multi-step workflows. Read more: https://lnkd.in/gaVVdi9H

🧩 Cockpit — Tech Radar Prompt Expanding a codebase always adds entropy. If there’s one thing I’ve learned from Lucas Silva, it’s this: you can always be more productive and precise when you’re only developing what’s truly needed. In my early experience as a manager, at some point, our system became a museum of frameworks and “temporary” patterns. The cure was a single question I kept asking: what’s truly core? That became our anchor, our intellectual rigor to control entropy. We didn’t always agree, but that tension kept us sharp. Three questions we used to reset focus: • What’s becoming noise, and what’s becoming core? • Are we solving today’s problem or chasing adjacent ones? • As a technical leader, am I curating or collecting technologies? We logged every tool in a Tech Radar: Adopt | Hold | Watch | Retire, and bought ourselves back weeks of clarity. 💬 How do you separate noise from core in your stack? #TechRadar #Focus #Engineering #Productivity #Scaling

AI agents are evolving fast, and the next leap is all about collaboration. In our latest blog, Igor Brito, Senior Software Engineer at Cheesecake Labs, explains how MCP (Model Context Protocol) and A2A (Agent-to-Agent Protocol) are transforming the way autonomous systems connect, execute, and cooperate across domains. Learn how these open standards are paving the way for interoperable, multi-agent ecosystems that will redefine how businesses automate, decide, and scale. 👉 Read the full article: https://lnkd.in/dUrtrn7V

Every engineering team wants to move faster, but debugging in the field, chasing crashes, and managing returns slows everything down. Memfault helps teams launch and maintain reliable connected products with less friction. Our device observability and OTA platform automatically surface real issues, so engineers can fix bugs quickly and get back to building what’s next. As reMarkable’s CTO put it, integrating Memfault freed up 5–10% of engineering capacity — time they can now focus on innovation rather than firefighting.