Puneet Jindal

🚨 India’s consumption puzzle: Why do two official datasets tell different stories about how we spend? PFCE (top-down; macro-consumption) vs. HCES (bottom-up; micro-consumption) 🔹 Gap has widened from 5% in 1972–73 to ~45% today. 🔹 Free/subsidised services (education, healthcare, meals) inflate PFCE but barely appear in HCES 🔹 Food share in household budgets is falling; non-food (transport, healthcare, rent, education) dominates This divergence isn’t just statistical—it affects how we measure poverty, prosperity, and the impact of policy. 📊 Latest piece by Ashish Kumar and Payal Seth from Pahlé India Foundation (PIF) in businessline unpacks the data divide and calls for urgent reforms (https://lnkd.in/dJWjyCQe)

179

8 Comments

Introducing DynaRoute - intelligent model routing for efficient LLM usage At Vizuara, we (Pritam Kudale, Pranavodhayam P R, Raj, Rajat) have spent the last few months studying how teams use large language models in production. What we found was simple but eye-opening. Most teams rely on one large and expensive model for every kind of query, whether it is a quick factual question or a complex reasoning task. This works, but it wastes a lot of money and computing power. That is what led us to build DynaRoute (https://lnkd.in/dGSx2KRq), an intelligent routing layer that automatically analyzes your queries and sends them to the most suitable model based on complexity, cost, and performance. You do not have to change your application or rebuild your backend. You can continue using your familiar OpenAI-style API, just by switching the base URL and key. From that point, DynaRoute handles everything intelligently behind the scenes. In our internal benchmarks, DynaRoute delivered remarkable results. Real-world employee questions achieved up to 92 percent cost savings, while code-related queries saw around 74 percent savings. All of this happened while maintaining nearly the same quality and accuracy as top-tier models. For teams that are running customer support systems, coding assistants, agents, or RAG-based pipelines, DynaRoute can reduce your inference bills drastically while preserving the quality your users expect. You can think of it as a smart, adaptive router that continuously learns how to choose between models like GPT-4.1-nano, Gemini 2.5 Flash, and Claude Sonnet, depending on the query. We also provide API access so you can integrate DynaRoute directly into your own platform or agent framework. The API is OpenAI-compatible and can be configured with just a few lines of code. Here are our current plans: Free Tier: ₹0 per month, includes up to 50,000 tokens, OpenAI-compatible endpoint, and access to premium models. Plus Tier: ₹888.30 per month, includes up to 3 million tokens, unlimited requests, access to premium models, and priority support. Enterprise Tier: completely custom implementation for companies that need dedicated support, detailed analytics, and on-premise deployment. We also provide custom routing models for enterprise-level optimization. We are currently offering Enterprise Plan API access to selected partners who wish to integrate DynaRoute into their platforms. If you want to optimize the cost and performance of your LLM-based applications without compromising output quality, give DynaRoute a try. 👉 Learn more and get your API key at https://lnkd.in/gTrs9NWA

84

3 Comments

Recently, PW (Physics Wallah) launched Aryabhata 1.0, an AI model built using something called an SLM, which stands for Small Language Model. Unlike big models like GPT-4 or Gemini (called LLMs – Large Language Models) that need a lot of data, power, and memory to run, SLMs are smaller, faster, and easier to use. They are made for specific tasks and give quick answers without using too many resources. This makes them a good fit for education, where speed and accuracy are more important than handling every type of question. While LLMs try to do everything, SLMs are focused — perfect for things like doubt-solving, tutoring, and exam help. Note: A lot of news has surfaced claiming that it is just a wrapper on top of the Qwen model. However,this post is solely focused on how LLM and SLM models differ.

730

8 Comments

India has only 14 pathologists per million people. And they’re manually reviewing every diagnostic report. That’s thousands of reports a day, per expert. No surprise—studies show 5% of reports contain errors. Now imagine this: An AI agent trained to scan every report. It flags only the doubtful ones for the pathologist. Burden reduced by 90%. Accuracy improved. Speed multiplied. This is not the future. We’re already doing this at Flabs. But here’s the catch— 90% of diagnostic labs in India don’t even have a digital application layer. Forget AI—they’re stuck in the paper era. We’re not building Flabs just for today. We’re building it for the India of 2035. Where AI works alongside doctors. Enjoy the the video. #VerticalAI #Diagnostics #HealthTech #SaaS

50

4 Comments

When hiring AI talent, the core skills you want to assess fall into four main categories: Technical depth, Problem-Solving, Product Impact, and Collaboration. Here’s a breakdown of the most critical skills to evaluate: Technical Skills ➕ Machine Learning & Deep Learning Foundations -Supervised/unsupervised learning, overfitting, regularization, etc. -Deep learning architectures (CNNs, RNNs, Transformers) Applied MLOps / Production Engineering -Model deployment, inference optimization (ONNX, TensorRT, vLLM) -CI/CD for ML, versioning (MLflow, DVC), containerization Programming & Data Manipulation -Proficiency in Python and libraries like NumPy, Pandas, PyTorch, TensorFlow -Data engineering pipelines (SQL, Spark, Airflow) LLMs and GenAI (if relevant) -Prompt engineering, fine-tuning, RAG pipelines, LoRA -Knowledge of vector databases (e.g., FAISS, Pinecone), embedding models Mathematical & Statistical Reasoning -Probability, linear algebra, optimization, and statistics The ability to break down complex business or research problems and the creativity in applying the right algorithm or model to a use case is everything. Don't be to lost in the hype, stick to the fundamentals! #oryxsearch #AI

120

6 Comments

Python Curriculum at Deerwalk Schools – Preparing Kids for the AI Era At Deerwalk, we have been teaching coding since Grade 1. Now, we are making it more structured from Grade 3 onward and putting stronger focus on Python—the world’s #1 language for AI, data science, and automation. Why Python? Because coding is today’s literacy. Just as students must learn to read, write, and calculate, they must also learn to code. Python is simple enough for a 10-year-old to start, but powerful enough to run Google, NASA, and ChatGPT. By the time Deerwalk students reach Grade 10, they will be fluent coders, ready to apply Python in real-world projects. We are sharing our roadmap below not just for Deerwalk parents and teachers, but also for other schools in Nepal who want to follow Deerwalk’s lead. Python Roadmap at Deerwalk Schools Grade 3–4: Logic & Fun with Code Block-based coding (Scratch, Blockly) Transition to Python with Turtle graphics Projects: Animated stories, drawing a house with Python Grade 5: First Steps in Python Variables, input/output, Turtle shapes Projects: Guess-the-number game, digital dice roller Grade 6: Building Foundations Loops, conditionals, random numbers Lists and simple data storage Projects: Rock-Paper-Scissors, simple chatbot Grade 7: Expanding Skills Functions and modular code Dictionaries, file handling basics Projects: Word counter, Hangman game Grade 8: Problem Solving Search & sort algorithms Error handling, debugging Projects: Student grade tracker, text adventure game Grade 9: Applied Python OOP basics Data visualization with Matplotlib APIs and real-world data Projects: Weather app, expense tracker Grade 10: Python Fluency Web basics (Flask) Data analysis with Pandas Intro to AI/ML (Scikit-learn) Projects: Sentiment analyzer, quiz web app Grade 11: Python for Innovation Advanced OOP & design patterns Machine learning projects Robotics with Python (Raspberry Pi/Arduino) Projects: Handwriting recognition, smart home simulation Grade 12: Python for the Future Deep Learning (TensorFlow/PyTorch) Capstone projects solving real-world problems Projects: AI chatbot, recommendation engine Why This Matters By following this roadmap, Deerwalk students will graduate with 10 years of structured coding experience, fluency in Python, and confidence to build AI-powered applications before leaving high school. This is how Deerwalk is preparing the next generation. We invite other schools to join us in giving kids the real skills of the future.

88

5 Comments

Some people crack AI/ML roles at Google, Microsoft, Apple on their very first attempt. But nobody tells you this secret 👇 It’s not luck. It’s a method — and anyone who follows it with focus and consistency can do the same. 🔹 The “Math & ML First” Principle Before touching deep learning frameworks, they master the core: • Linear Algebra & Calculus • Probability & Statistics • ML Algorithms: SVM, KNN, Decision Trees • Optimization (SGD, Regularization) Because interviewers don’t just ask you to use models — they ask why and how they work. 🔹 The “Concepts Over Code” Rule They don’t rush to write 10,000 lines of TensorFlow or PyTorch. They first understand bias-variance tradeoff, overfitting, feature engineering, and model evaluation. Then they code. 🔹 The “Projects That Matter” Approach Instead of copying 10 Kaggle notebooks, they build 2 solid end-to-end projects: • With real datasets • Using data cleaning + ML pipelines • APIs, dashboards, or model deployment That’s what hiring teams remember. 🔹 The “Paper + GitHub” Combo They read 2–3 core ML/DL research papers — and replicate or extend them on GitHub. This shows real depth. Not just course knowledge. 🔹 The “Mock & Explain” Mindset They practice mock interviews — not just solving, but explaining: • Why they chose this model • Why accuracy dropped • Why this feature matters Because AI interviews test your thought process, not just code. If you're serious about cracking top ML/AI roles — Stop chasing random tutorials. Start following a proven strategy that works. Say “MLMETHOD” and I’ll share the exact roadmap followed by folks who made it to Google Research, Razorpay AI, Microsoft Applied Sciences, and more. Someday, someone will say: “You just got lucky.” But you’ll know — you followed the system nobody talks about. Please Check Best 💥 Discount 💪 Products 🌐 https://lnkd.in/gfZJqPTR Follow ✅ https://lnkd.in/gZ6myPbn Please follow Chanchal Kumar Mandal #MachineLearning #AIJobs #MLInterviewPrep #TechCareers #SDEPrep

658

2 Comments