Delegation and Agentic for Everyone
A weekly recap — why delegating to AI agents is the same skill managers already learned, and why an agent's invisible work is what makes it so hard to grasp for people who don't code.
Sidharth Ramachandran Blog
AI systems, product strategy, and applied machine learning.
Latest posts
Thinking Tokens & Doing Tokens: Agents and Harnesses
A weekly recap — untangling what a harness actually is, and adding agentic monitoring to Podsumo while learning to orchestrate with Fable.
AI's Little Secret: We've Been Doing Testing All Along
A talk at QonfX Berlin on how traditional testing practices are converging with AI evals — and why the two disciplines have more in common than either side realises.
Understanding Speculative Decoding
A practical explanation of why speculative decoding speeds up generation in large language models.
The EU Is Trying to Regulate AI. But First, It Has to Define It.
A practical read on general-purpose AI definitions and why regulatory clarity matters.
Scaling asset generation using Generative AI
How generative systems can act like a production line for creative assets, with humans in the loop.
Human-AI collaboration and understanding AGI
From task-specific agents to general intelligence: practical reflections on hardware, software, and AI assistants.
The amazing journey of text classification
A decade-long perspective on how text classification evolved from lexicon methods to LLM prompting.
My personal experience with ChatGPT and a vision for the LLM future
Reflections on using LLMs in healthcare conversations and the future of personal digital assistants.
Encoding in Spark vs. Pandas
How Spark and Pandas handle One Hot Encoding differently — dense vs. sparse representation — and a practical workaround for interoperability.
My experience at Techcrunch Disrupt
A first-hand account of attending TechCrunch Disrupt Berlin 2017, building a WhatsApp chat assistant overnight, bombing the demo, and somehow ending up featured on TechCrunch.
Data Science Recruiting — Part 2
A walkthrough of the common recruitment process steps I've observed for data science teams, from resume screening through face-to-face interviews.
Recruiting for Data Science Teams
Key parameters for evaluating candidates when building a data science team, from educational qualifications and projects to communication and attitude.