So you built a RAG system. You watched one YouTube tutorial, copy-pasted some LangChain code, threw 3 PDFs at a local Chroma instance, and it answered "What is the refund policy?" with suspicious accuracy.

You showed it to your manager. They clapped. Someone said the word "production." And now you're here, because production is not a demo with more users. Production is a demo that has been sleep-deprived, underfed, and actively lied to by your own data pipeline.

By now “RAG” made you roll your eyes. But now we’ve got MCP—Model Context Protocol. Yeah, yet another acronym to make devs sound important. Here’s the deal: it's basically a universal way for AI apps to plug into your world—your files, tools, databases—without custom duct tape between systems. 😅

So, you’ve heard the word RAG. Sounds fancy, right? Don’t worry, it’s just another AI buzzword designed to make engineers feel important at meetups. But here’s the kicker: plain RAG works… until it doesn’t. Enter rerankers, the secret sauce that makes your “meh” RAG pipeline slightly less embarrassing. 🙄

Yeah, sounds like the latest buzzword from the AI hype machine, right? Whoooo hyped term haan!!!. But let’s pause the FOMO and actually break this thing down into bite-sized chunks. 🙄 Let’s rewind to basics before diving into the alphabet soup.

Welcome to the land of AI buzzwords. If you’ve survived RAG, MCP, and “AI Agents,” congratulations—your buzzword tolerance is above average.

But let’s go deeper. Ever wondered how a thing like ChatGPT or Claude or Gemini or DeepSeek actually gets made? You hear numbers like 7B, 70B, 1 trillion parameters—but what does that even mean? Why does Anthropic call their thing “Constitutional AI”? Why is DeepSeek suddenly everywhere with “cheap inference”? And most importantly, why do these model files weigh more than your entire Steam library?

Ah yes, Dijkstra’s algorithm. The one you swear you’ll never forget, and yet somehow in every contest you end up trying BFS on weighted edges (“because it worked last time, right?”). Spoiler alert: BFS is not magic. If edges have weights other than 1, you’ll get the wrong answer faster than your WA verdict appears. So buckle up, because we’re about to roast your brute force instincts while learning to drive this algorithm properly.

Ah yes, SOLID principles — the holy commandments every senior developer swears by. Single Responsibility, Open-Closed, Liskov Substitution, Interface Segregation, Dependency Inversion. You’ve probably heard your architect chanting them in daily stand-ups like they’re sacred verses. 🤓

But here’s the spicy truth nobody tells you: these principles do make your CPU cry and your RAM drink Red Bull. 😎

Shocked? Oh, you thought your precious IUserService was free? Cute.

So, you’re a developer. You’ve been told polymorphism is this magical wand that makes your code flexible, extensible, and architecturally pure.

But let’s be honest — nobody told you it also makes your CPU sweat like it’s running a marathon in Bangalore summer. 😂

Welcome to the dark side: understanding the overhead of polymorphism.

This blog covers analyse of the Weather data-set of Finland, a country in the Northern Europe. Recorded for last 10 years using Python. It covers basics of Pandas and NumPy libraries for data cleaning and preprocessing part and Matplotlib for presentation of results.