360SOFTY

Insights

Engineering Insights

Practical writing on software architecture, SaaS products, AI automation, legacy modernisation, and the business of building reliable systems.

RSS

Curated links from external sources — not 360Softy original articles.

ExternalSoftware Engineering
DEV Community

Foundation vs. Instruct vs. Chat Models: One Question, Three Answers

A hands-on tutorial you can run for free in Google Colab. Run it yourself: open foundation_instruct_chat_tutorial.ipynb in Google Colab and run every cell top to bottom. It uses the SmolLM2-135M family — small enough for a free CPU runtime, no GPU needed. People say "LLM," "GPT," "an AI model," and "ChatGPT" as if they were the same thing. They aren't. There's a ladder of training stages between "a model that read the internet" and "an assistant you can chat with," and the words foundation, inst

aibeginnersllm
DEV CommunityRead original
ExternalSoftware Engineering
DEV Community

We built an app for Japanese neighborhood associations — and the handover notebook turned out to matter most

The problem Every year in Japan, a small group of volunteers runs the local 自治会 (neighborhood association). They circulate bulletins, collect dues, check in on elderly residents, run votes at general meetings, and maintain the annual schedule. At year's end, they hand the role over — usually with a folder of paper, or nothing at all. Nobody tracks who did what. Nobody measures the actual workload. So when it's time to recruit the next round of officers, no one can honestly answer "how much wor

productivityiosflutter
DEV CommunityRead original
ExternalSoftware Engineering
DEV Community

Product Launch Telemetry: How AIKit Turns Feature Releases Into Agent-Readable Growth Loops

Product launch telemetry is the bridge between shipping a feature and proving that the feature created growth. AIKit can turn every release into an agent-readable loop by publishing the launch page, instrumenting CTA events, summarizing customer signals, and feeding the next experiment back into the roadmap. Most product launches create a burst of activity and then lose the evidence. A team ships a feature, posts an announcement, watches a few dashboard numbers, and moves on. Two weeks later nob

webdevproductivitytutorial
DEV CommunityRead original
ExternalSoftware Engineering
DEV Community

Zero Variance: Proving COBOL-to-Java Semantic Equivalence with a Live Mainframe Emulator on AWS

By Banu Parasuraman Distinguished Engineer | Account CTO | Mphasis "The mainframe isn't going away. But the teams who understand it are. The window to modernize safely — before institutional knowledge walks out the door — is closing." Mainframe modernization is one of the most consequential — and most feared — programs in enterprise IT. Billions of dollars of financial transactions flow daily through COBOL batch jobs that have been running, largely unchanged, for three to four decades. The fear

mainframeprogrammingmodernization
DEV CommunityRead original
ExternalSoftware Engineering
DEV Community

I'm building CortexDB — an agent-native context database for AI agents

I'm building CortexDB — an agent-native context database for AI agents Most modern RAG systems work like this: Split documents into chunks Generate embeddings Store them in a vector database Retrieve top-k similar chunks on query Send them to an LLM It works for simple use cases. But as AI agents become more autonomous and complex, a clear problem appears: Agents don’t just need similar text chunks. They need bounded, permission-safe, evidence-aware, and verifiable context. This is why I

ragagentsrust
DEV CommunityRead original
ExternalSoftware Engineering
DEV Community

Why AI coding agents need a launch layer

The coding got 10x faster. The launching didn't. Point a coding agent at a half-formed idea and it will hand you a working app before your coffee is cold. Backend, frontend, schema, a couple of API integrations, tests that pass. The part that used to take a week of focused work now takes an afternoon of prompting. That much is real, and you've felt it. Then you try to ship the thing, and time stops moving. Because the agent didn't write the part that turns code into a product. It wrote the busin

aiagentsdevtools
DEV CommunityRead original
ExternalSoftware Engineering
DEV Community

Why AI agents can't shop today — and what we're building to change that

AI agents are getting smarter every week. They can browse the web, write code, analyze data. But there's a glaring gap in their capabilities: they can't interact with real commerce. Ask your favorite AI agent "what's the cheapest 5kg bag of rice I can buy in Lima right now?" and watch it struggle. It doesn't have access to real-time retail prices. It doesn't know which stores carry rice. It can't compare prices, let alone complete a purchase. This isn't a model capability problem. It's an infras

aiarchitectureapi
DEV CommunityRead original
ExternalFrontend Development
Vercel Blog

Vercel for Enterprise Apps and Agents

Today we are introducing , a platform that gives your entire company the ability to ship with AI safely, behind your access and security boundaries.Vercel for Enterprise Apps and Agents Over the past year, employees across Vercel shipped hundreds of agents and internal apps. Getting to production was the easy part, because we built them with on top of the and deployed them on Vercel.eveAgent Stack The difficult questions came after those agents were being used by our employees across the compa

Vercel BlogRead original

Work with 360Softy

Building a SaaS product, AI system, or business platform?

Book a free consultation and we will tell you honestly whether we can help.

Book a consultationView our services →