Kumar Sanu

🛠️ Backend Engineer | Architecting scalable systems & managing deployments 🚀 | Now diving deep into AI & LLMs 🤖 | Learning, building, sharing.

[MisogiAI @misogi_ai]

Imagine asking a simple question… and getting back not just an answer, but a mini-lesson in Hinglish, with voice-over, visuals, and even a quick quiz to test yourself.

This is what @misogi_ai students have built.
Try it out : https://t.co/OMhL1ehYtj

@shanu_1012 @Venkatalokesh0

#aiprojects #hirings #chatgpt #aitools

[Kumar Sanu @shanu_1012]

📄 Upload resume once → 💼 Jobs + 📝 Cover letters → 📲 Telegram notifications.
Job hunting, automated.
⚡ #Automation #JobSearch #n8n #AIAgent

[Kumar Sanu @shanu_1012]

🚀 Brainstorming at Misogi HQ! 📚
Diving deep into Attention Is All You Need — transformers, multi-head attention & positional encodings all over the board.
Innovation starts with whiteboards like these. ⚡️

#AI #DeepLearning #Transformers #Misogi

[Kumar Sanu @shanu_1012]

🎯 Set temperature = 0 and the model becomes a precision machine.
Ask it twice, get a carbon copy response — every word, every comma, identical.

[Kumar Sanu @shanu_1012]

LLMs and AI are cool.

But they become unstoppable when picked up by a curious software engineer who’s obsessed with solving, scaling, and shipping. 💻⚡

That’s when the real magic happens.

#AI #LLM #SoftwareEngineering #BuilderMindset

[Kumar Sanu @shanu_1012]

LLMs + RAG + basic caching ≠ scalable AI 🧠

Redis vector search + aliasing =
⚡ Smarter cache
📉 98% fewer LLM calls
⏱️ <1ms response

🔗 Full breakdown in pic👇
https://t.co/2sKiVtabBL

#LLM #RAG #Redis #VectorSearch

[Kumar Sanu @shanu_1012]

🚀 I built a CLI tool to compare LLMs across providers:
🔹 Base vs Instruct vs Fine-tuned
🔹 OpenAI, Hugging Face, Anthropic, Local
🔹 Same prompt, 4 different brains 🧠

Biggest lesson?
🎯 The smartest model isn’t always the biggest — it’s the right one for the task.

#AI #LLMs

[Kumar Sanu @shanu_1012]

@_aman_vats