⚡️ 0.35 to 0.81 in three sprints. No model changes, just better infrastructure. We just wrote the guide on building observable, reliable #AIAgents in production. Read more >> nf.ie/3NBHJuZ #AINativeEngineering #AIEngineering #Mastra

From AI prototype to production: how to build evals for reliable agents | Nearform

We're sharing the principles and processes that help us deploy resilient agents to production faster

nf.ie

Claude Code Teams vs Gastown - what they actually look like, how they differ, and which to reach for first. Our Head of AI, Cian Clarke, show us in under 10 mins. 👇 youtu.be/vQkJaM2NZPc #AIEngineering #AgenticAI #ClaudeCode

Claude code teams vs Gastown: Multi-Agent orchestration compared (under 10 minutes)

YouTube video by Nearform

youtu.be

Most AI talks give you the theory. This one gets right into the stack. Future Form is back in London on 10th March. 📆 Details and register info >> luma.com/q7sjimdw #TechMeetup #FutureForm #LondonTech #TechEvents #Nearform

If you think AI makes product thinking optional, it doesn’t. Our senior product designer built a Strava-connected web app in 3 days using BMAD, Cursor, and Claude. The difference wasn’t “coding”. It was framing the problem, asking better questions, and designing for real users. → nf.ie/4s5uj8S

On 25 Feb, Katie Roberts joins Dave Kerr (QuantumBlack AI by McKinsey), Cian Clarke and Hélène Haughney to explore how enterprises are adapting engineering practices to scale AI safely and sustainably. If you’re moving beyond pilots and into delivery, this conversation is for you. >> nf.ie/4qlTweZ

On February 25th, Cian Clarke joins Dave Kerr (QuantumBlack, AI by McKinsey), Katie Roberts and Helene Haughney of Nearform to share how enterprises are rethinking software delivery for the AI era, without increasing risk or creating AI debt. Register today >> nf.ie/4s0ZzGh

Copilot boosts productivity. It won’t scale delivery. Our next Impact Talks webinar brings together McKinsey and Nearform experts to explore what it really takes to industrialise AI in the enterprise. Move from just tools to a scalable, repeatable approach. Register >> nf.ie/4qlTweZ

We built a working demo of browser-based vector search using plain JavaScript - handling semantic similarity and filtering with no backend database. @ryan-roemer.bsky.social walks through the approach, data strategies, and performance results, showing how far on-device search can go. nf.ie/4rhtxWA

AI hype is peaking, but where's the real value? Nearform's Head of AI, Cian Clarke, breaks it down in Business Quarter: Ditch "plug-and-play" pilots for purposeful AI on strong data foundations and ops. businessquarter.co.uk/beyond-the-a...

Beyond the AI bubble — from hype to lasting impact

AI’s business promise is cooling as leaders confront implementation reality. Cian Clarke, Head of AI at Nearform, argues organisations must move beyond hype, rethink model strategy, and build durab…

businessquarter.co.uk

Too many teams are still relying on “vibe coding” when building with LLMs - jumping straight into prompts without clear requirements or long-term structure. Nearform’s Head of AI, Cian Clarke, breaks down how Spec-Driven Development offers a better way forward. tessl.io/blog/from-vi...

From Vibe Coding to Spec-Driven Development

Spec-driven development beats vibe coding by front-loading clarity, improving AI outputs, reducing churn, and scaling engineering workflows.

tessl.io

The model doesn’t just read your spec; it makes decisions when there’s ambiguity. Cian Clarke, Head of AI at Nearform joins @sjmaple.bsky.social from @tessl.io to explain why BMAD and spec-first dev are the future. 🎧 Watch now 👇 youtube.com/watch?v=WpH3...

Why Faster AI Development Often Increases Rework | Cian Clarke

YouTube video by AI Native Dev

youtube.com

The AI gold rush is over. Now it’s about adoption, accountability, and results. Our latest insight explores how 2026 will be defined by practical deployment, evolving engineering roles, and business models built for value, not volume. Read more >> nearform.com/insights/202...

No stand-ups. No deploys. Just downtime. 🎄 Happy Holidays from all of us at Nearform.

LLMs are only as smart as their context, something most teams realise too late. We break down the Model Context Protocol (MCP), a proven way to: - Structure data for LLMs at scale - Improve reliability and explainability - Avoid costly mistakes in production 📖 >> nearform.com/digital-comm...

For developers, using AI has moved beyond mere speed, it’s now about strategic planning. That’s why Alfonso Graziano tested Plan Mode in Cursor and GitHub Copilot, not just to see if they could 'vibe-code' a feature, but to determine: could they actually plan it?" 📖 Find out more >> nf.ie/44GZ4aQ

Cloud LLMs have their place, but on-device GenAI is fast, private and cost-free to scale. We built a working mobile demo that runs locally. No cloud, no API, full AI functionality. Explore the article + repo >> nf.ie/4rrhUws

AI is transforming design. But if we forget the human, we lose what matters most. Empathy isn’t optional. It’s the foundation of trust, relevance, and genuine innovation. Our Head of CX, Samantha Sherman, shares how to design for human connection in an AI-driven world. Read more below 👇

Designing for human connection in an AI world | Nearform

Nearform is an independent team of engineers, designers and strategists. We build digital capability and software solutions for ambitious enterprises seeking sustained business impact. We love what we...

nf.ie

Most teams say they’re “doing AI.” Few can show business impact. Our AI Native Engineering Field Manual was written by engineers, for engineers. Practical. Proven. Ready to use. Download your copy today. Find your first AI win. 📥 >> nf.ie/3Jdk9lS

If you’ve been wondering whether it’s worth switching from #Terraform, or just curious how #Pulumi handles at scale, read on here >> nf.ie/4ozlD9a 👇 Catch up on the whole series: Part 1 – Landscape analysis >> nf.ie/4osaK9K Part 2 – From chaos to clarity >> nf.ie/4giVSqx

The AI Native Engineering Field Manual is here. A practical guide for teams ready to go beyond AI pilots and build systems that learn, adapt, and scale. 🧭 Start your expedition into AI-native delivery. Free download here >> nf.ie/3Jdk9lS

Most #JS text chunkers are too heavy, too rigid, or too limited. So we built a small, fast, and smart alternative: llm splitter — designed for performance, extensibility, and storage efficiency. #Opensource and ready to use! 📖 Read the latest from @ryan-roemer.bsky.social >> nf.ie/439Lvjt

Testing can’t be an afterthought when models drive decisions. As AI becomes an integral part of critical infrastructure, testing needs to evolve accordingly. Our latest article examines how to test AI/ML systems that adapt and evolve without compromising safety or speed 👇 ➡️ nf.ie/4pYBu2G

Testing AI requires a new mindset. At #WMF2025 we shared how Nearform approaches LLM testing: - 5 levels of testing - Metrics & tools that work - Why hallucinations are bugs, not features Read the full article >> nf.ie/3VCqn1m

Meet Chekhov - our new #opensource tool that lets you chat with AI personas trained on real user data. → Clearer insight → Better product-market fit → Smarter decisions, sooner Read more ↓ nearform.com/digital-comm... #AI #product #LLMs #ux #devtools

Part 2 of our series is now live and explores how Pulumi performs in real-world scenarios. 📖 Read the full article → nf.ie/4giVSqx Part 1 here → nf.ie/4osaK9K for context on why we chose Pulumi and how it compares to alternatives in 2025.

Our many years of experience in building production-ready systems with #opensource technologies mean we can help: ✔️ Select and tailor the right models for your business objectives ✔️ Deploy with secure, reliable pipelines ✔️ Keep costs and long-term maintainability at the forefront nf.ie/4fwyeGM

AI can’t fix broken systems, it can just scale them faster. James Malone explains why role-based AI agents and solid engineering matter more than ever. 👉 nf.ie/41UZRUb

OpenAI’s #opensource reasoning models aren’t just about using tools, they’re about improving them. Colm Harte & Sarah Egan share why contribution matters: ✔️ Improving docs & sharing test cases ✔️ Building new tooling & modules ✔️ Strengthening the #OpenAI ecosystem Learn more → nf.ie/4fwyeGM

Most product teams skip personas not because they don’t care, but because they don’t have time. So we built Chekhov. ✔️ Upload product docs ✔️ Chat with dynamic, AI-generated personas ✔️ Built with #OpenAI + Replit See it in action → nf.ie/41R7aMC #AI #ProductDesign #UX #Nearform #LLMTools

Open source AI gives you control but it also gives you responsibility. Nearform’s Colm Harte & Sarah Egan explain: ✔️ Security, scalability, and monitoring requirements ✔️ The DevOps and governance work you’ll own ✔️ Why the right partner can make all the difference Full article >> nf.ie/4fwyeGM