AI TCH RA Rahultiwari I Explored Open-Source Models with Ollama as Alternatives to Claude-Based Workflows Are open-source models ready to replace Claude? Not directly. But they are becoming very useful.
MDA RA Rahultiwari From Transcription to Reasoning: What Audio Flamingo 3 Changes in an Audio Stack I still remember the moment I realized I’d been treating audio like a bolt-on to my AI stack. And it was costing me.
AI SCI RA Rahultiwari Synthetic Data Finally Made Sense to Me When I Stopped Thinking of It as “Fake Data” For a long time, the phrase synthetic data sounded slightly strange to me. We spend so much time in data science trying to get real data.
HUM RA Rahultiwari How I Started Understanding NLP: Turning Messy Text Into Useful Signals When I first started learning Natural Language Processing, I thought it was mostly about making machines “understand language.”
SPT AI RA Rahultiwari Stemming vs Lemmatization in NLP: What I Understood After Actually Using Them in Text Preprocessing I have been working around NLP preprocessing for a while now, and one thing that looks very small in theory but becomes important in…
AI HUM RA Rahultiwari Cross-Validation in Machine Learning: The Simple Idea That Took Me Some Time to Really Understand When I first learned about model validation, I thought it was simple.
AI SCI RA Rahultiwari I Looked at Karpathy’s AutoResearch, and It Changed How I Think About ML Experimentation I spend most of my time in the reality of AI research: models, ablations, evaluation, failure cases, and the long stretch between an idea…
AI RA Rahultiwari GPT-6 Astra: We Are Moving From AI That Answers to AI That Actually Does the Work OpenAI’s latest model feels less like another chatbot upgrade and more like a preview of where agentic AI is heading.