Kashyap M Kale
@kashyap_ai
Followers
27
Following
869
Media
8
Statuses
121
Software Engineer @ Tray | Ex-Amazon | Algorithms | Leetcode | Artificial Intelligence | ML | Backend Engineering
Hyderabad
Joined June 2024
F*ck cold email cold LinkedIn DM outreach is the new way we’re getting 10% reply rates, on our worst campaigns i made a full guide going over our EXACT system A-Z RT + reply "system" and i'll send you the guide (make sure you're following so I can DM)
229
114
208
- every mistake is a lesson - but every risk is a possibility Good one, read somewhere.
0
0
0
I could be wrong, but it seems like traditional software engineering roles may become completely irrelevant in the coming years. However, skilled computer engineers those with a strong grasp of systems, hardware, and low-level computing will likely remain in high demand. #AI
0
0
1
Cutting the noise. 🎯 Too many priorities = zero progress. Time to fix that. Current 📊 ✅ 86/150 NeetCode ✅ 3/15 System Design (Alex Xu) Target by July 31: 🚀 112/150 NeetCode 🚀 15/15 System Design Let's go! 💪#SystemDesign #LeetCode #LearnInPublic #FocusMode #buildinpublic
0
0
5
8/8 This highlights the importance of understanding each model’s strengths when choosing the right AI tool for your specific needs. #AI #PromptEngineering #ChatGPT #Gemini #MachineLearning #AIInsights #LLM
0
0
0
7/8 What are your thoughts on this behavior? Have you observed similar patterns in model performance and prompt sensitivity?
1
0
0
6/8 🤔 Key questions: • Is ChatGPT’s transformer architecture better optimized for intent recognition? • Are other models using more conservative inference strategies to avoid hallucinations? • Do they have different optimization objectives?
1
0
0
5/8 Other LLMs often need much more explicit instructions and structured input to deliver similar results. This suggests differences in: • Training data processing • Attention mechanisms• Fine-tuning approaches
1
0
0
4/8 This isn’t a one-off! I’ve been noticing this pattern repeatedly across different use cases. ChatGPT demonstrates superior contextual inference and requires less detailed prompts to achieve desired outcomes.
1
0
0
3/8 ChatGPT’s interpretation: ✅ Got it right - created an email to Tripti to find the count of furniture in the email. Gemini’s interpretation: 🔄 Different angle - wrote an email asking Tripti to help count chairs and tables, treating it as a request for assistance.
1
0
0
2/8 What I expected: An email to Tripti listing the count of furniture items that would be included below in the email.
1
0
0
🧵 I’ve been testing various AI models with the same vague prompt, and the results are fascinating! The prompt: “Write an email to Tripti saying please find the count of chairs and tables below in a sweet and human tone” #LLM #AI #AIEngineering #GenAI
1
0
0
Twitter isn’t social media. It’s a battlefield. Open the app: — 18 y/o building a startup — Someone just made $100K — Another grinding in silence at midnight It’s not Insta’s dopamine trap. It’s a mirror. Either you level up or get left behind.
25
22
279
🔧 Stack: React, API Gateway, Lambda, S3, MySQL, Amazon Bedrock, FAISS. 🚀 Future: Enhanced UX, auth, scalability! #AIEngineering #AWS #Innovation #Lambda #CloudArchitecture #Startup #TechStartup #SaaS #AIStartup #Founders #BuildInPublic
#LeanStartup #GenerativeAI #RAG
0
0
3
🧠 Cost-optimized design: RAG + FAISS avoids sending full transcripts to LLM per query. Only relevant chunks used → Slashes token costs! AWS Lambda’s pay-per-use = scalable + budget-friendly. #CostOptimization #LLM
1
0
0
👉 Demo: https://t.co/Y5p2q9xetj 💡 Architecture Diagram Below: Handles queries like: • “When are TA office hours?” • “Key points in Lecture 3?” • “How to submit Project 2?” • “Explain AWS Lambda architecture slides?” #GenAI
1
0
0
🚀I saw a common problem at most grad schools, TAs are always loaded with repetitive questions even after clear announcements. So I built an AI Lecture Assistant in just 1 week as an assignment for my Cloud Computing Course ! #BuildInPublic #GenerativeAI #RAG #AIEngineering
1
0
3
A man can only dream Mac M4 Pro A100 GPU workstation Dual 4K monitors Cursor Pro Claude Opus + GPT-4 White Monster Razer keyboard Noise-canceling headphones
26
6
181
🚀 LeetCode Daily Problem Explanation for July 10th! 🚀 #LeetCode #Coding #Programming #Algorithm #ProblemSolving #Tech #Developer #CodeNewbie #100DaysOfCode #Interviews #DSA #SDE #TechLife #CodeChallenges #TechTips 💻🔥
0
0
1