We turn the unthinkable into reality by understanding our clients' challenges, envisioning how software can solve them, and delivering the right solutions.

London, England
Joined February 2011
Not every AI task needs a chatbot. Jev AI by TypeSafe returns typed decisions (Choice, Score, Noul) instead of text, built for routing, triage and branching logic. Our dev guide covers the API and how to test it before production: codescrum.com/blog/jev-ai-de… #AI #LLM #API
2
1
69
Local AI's biggest problem was never running a model — it was stitching together RAG, fine-tuning, APIs and tool use across five different apps. Unsloth Desktop is trying to fix that by putting it all in one workspace. Wrote about it here 👇 codescrum.com/blog/unsloth-d… #ai #llm
1
219
AI coding is moving past "prompt → generate → review → repeat." AI loops flip it: you define a goal + a trigger, the AI evaluates progress, acts, checks, and keeps going until the objective is actually met. No human re-prompting every step. New post 👇 codescrum.com/blog/ai-loops-…
184
Open-source AI coding agents are changing how developers build software. Our latest article explores OpenCode AI: ✅ Multi-model support ✅ Privacy-first architecture ✅ Agent Skills ✅ Open-source flexibility Read now👇 codescrum.com/blog/opencode-… #OpenSource #AI #DevTools #Coding
237
3 engineers. 1M lines of production code. 5 months. Zero written manually. That's what OpenAI pulled off with harness engineering — the methodology every CTO needs to understand in 2026. It's the next evolution 🧵👉 codescrum.com/blog/what-is-h…
41
AI agents don't just chat — they act. We broke down OpenClaw, the viral agent redefining autonomous automation 🤖→ codescrum.com/blog/openclaw-… #AIAgents #Automation
1
36
What if AI could actually USE your computer, not just answer questions? That's exactly what's happening with AI agents and Claude's computer use API. I explored the agentic revolution and what it means for how we work: codescrum.com/blog/the-rise-…
1
53
🧠 Beyond LLMs: World Models are the next AI breakthrough World models build internal simulations—predicting consequences, planning actions, understanding physics. Latest from @codescrum on why this matters 👇 codescrum.com/blog/world-mod… #AI #MachineLearning #WorldModels
37
The AI landscape is moving so fast that even experts can't keep up. We analysed 25 essential AI tools that will define professional excellence in 2026. Were's what you need to know! 🧵 codescrum.com/blog/the-ai-re… #ai #tools #GenerativeAI #futureofwork #work
44
Vibe coding is changing how we build software 🚀 Instead of writing code line by line, developers now describe what they want in plain language — and AI does the rest. Is this the future or are we losing something essential? Our deep dive 👇 codescrum.com/blog/vibe-codi…
44
🚀 New blog out: How to Reduce Non-Determinism & Hallucinations in LLMs Why does ChatGPT sometimes answer differently to the same prompt? And why does it “make stuff up”? Dive in👇 codescrum.com/blog/how-to-re… #AI #LLM #NLP #MachineLearning #TrustworthyAI
35
Two “open” AI drops in Aug 2025: ⚡ xAI’s Grok 2.5 (~270B params) ⚡ OpenAI’s GPT-OSS 20B & 120B But their take on openness couldn’t be more different. We compared them head-to-head—licensing, transparency & EU AI Act impact. 👇 codescrum.com/blog/xais-grok…
102
Just read “LiveKit Explained: The Open-Source Framework for Real-Time Audio and Video" here: codescrum.com/blog/livekit-e…! LiveKit leverages WebRTC + SFU to deliver low-latency, scalable audio/video. Worth a look if you’re building real-time apps!
1
69
Just read this 👇 Runner-H & Surfer-H by H Company are pushing the limits of what AI agents can do on the web. Autonomous search, analysis & action. Wild stuff. 🔗 codescrum.com/blog/runner-h-… #AI #AutonomousAgents #WebAutomation #FutureTech
57
AlphaEvolve by @DeepMindAI is an LLM-powered evolutionary agent that writes and optimises its own code — beating a 56-year-old algorithm & speeding up LLM training by 23% for a 1% cut in training time! 🧠🚀 Dive deeper: codescrum.com/alpha-evolve-b… #AI #DeepLearning #Code
1
73
🚀 Google’s Agent Space lets AI agents talk via A2A—seamless automation & productivity. See what it means for your business. Full story 👇 codescrum.com/blog/google-s-… #AI #AgentSpace #A2A
73
🚀 The future of AI is efficiency! In our latest monthly article, we explore how DeepSeek R1 is redefining AI with its groundbreaking efficiency and performance! codescrum.com/blog/the-futur… #AI #MachineLearning #DeepSeekR1 #TechInnovation
71