Developer Builds AI Cost Tool Where LLM Explains Decisions, Not Makes Them

·1 views

A developer building an Azure Cost Intelligence Platform discovered that AI-generated infrastructure recommendations often contained errors, including non-existent VM types and invalid CLI commands. To fix this, the architecture was redesigned so that independent components — including a metrics engine, pricing engine, and deterministic rule-based recommendation engine — gather and process real data before any AI is involved. The large language model is only used at the final step to explain pre-verified recommendations in plain language, never to generate them. The platform pulls live data from Azure Monitor, Azure Advisor, and Azure Pricing APIs, ensuring all suggestions are grounded in verified facts. The developer concluded that AI tools in cloud infrastructure are most reliable when they assist human understanding rather than drive automated decision-making.

Read the full story at DEV Community

This is an AI-generated summary. ShortSingh links to the original source for the complete article.

Discussion (0)

How One Word — 'Trade-off' — Defined a Computer Engineer's Entire Career

A computer engineering student in South Korea was introduced to the concept of 'trade-off' by Professor Sang Lyul Min during an introductory course that also served as his first exposure to C programming. With limited internet access at the time, the student had to look the term up in a Korean-English dictionary, gradually grasping its deeper meaning. The word came to encapsulate a core principle: that every engineering decision involves a compromise, distinguishing the discipline from pure science or mathematics. As the author progressed through his career — from algorithms courses weighing time against space complexity to reading academic papers — the concept of trade-off resurfaced repeatedly. Decades later, he considers it the single most defining word in engineering, one he was effectively handed for life on his first day of college.

0 comments Read more at DEV Community

ProgrammingDEV Community ·

How Apex Code Can Automate Gladly Task Creation Directly from Salesforce

Developers can use Salesforce Apex to integrate with Gladly's task API, automatically creating customer-linked follow-up tasks when events like appointment rescheduling occur. Gladly tasks are customer-specific records that sit on a customer's timeline alongside conversations and emails, requiring both task details and customer identification in the API request. The integration uses a wrapper class pattern in Apex to manage the nested JSON request body, making the code easier to maintain as the integration evolves. Gladly's POST endpoint accepts fields such as assignee, task body, due date, and customer contact information to create or attach tasks to existing customer profiles. Building in error handling from the start is emphasized as critical, since skipping it often leads to complications during real production incidents.

0 comments Read more at DEV Community

ProgrammingDEV Community ·

No Publishable Content Found in Submitted Article

The submitted article contains no factual or newsworthy information. The text consists solely of placeholder content described as a casual, inconsequential post. There are no identifiable subjects, events, locations, or dates present. As a result, no meaningful summary can be accurately produced from this source.

0 comments Read more at DEV Community

ProgrammingDEV Community ·

Developer builds custom Django command to alter Milvus schemas without data loss

A developer at a company using Milvus as its vector database needed to modify collection schemas in production without losing existing customer data. At the time, Milvus lacked a native ALTER command, with the only official workaround being a manual data migration to a new collection. To streamline this process, the developer wrapped the migration logic into a reusable Django management command that any team member could run safely. The command accepts a database name, collection name, new schema definition, and batch size, then creates a temporary collection, copies data in batches, drops the old collection, and renames the temporary one. Field removals and additions with default values are handled automatically, though updating existing field values is not currently supported.

0 comments Read more at DEV Community