Tanishq Singh·Sep 29How Andrew Ng’s OpenWorker Orchestrates ContextThe interesting part is not the usual model → tool → result loop. Most agent frameworks already do that.
Tanishq Singh·Aug 19How to write effective LLM as a Judge?I wrote earlier about what to test in a RAG system and how you can use LLM as a Judge. This post is about the harder part, which is writing…A response icon1A response icon1
Tanishq Singh·Aug 10What to Evaluate in a RAG Pipeline?Continuing from the last post. You know why you can’t hand off evals to someone else, and why error analysis has to come before you pick a…
Tanishq Singh·Aug 2Evals is QA for AI Engineers!A normal test checks an expected value. You know what a particular function should return, you write a test, check the return value!
Tanishq Singh·Jul 20Automated Red-Teaming : GPT-RedEvery agent that reads third-party data has an attack surface. A browser tab, an email body, a tool response, a line in a code repository…
Tanishq Singh·May 27ReAct: How LLMs Learned to Think Out Loud Between ActionsIn late 2022, two threads of LLM research were running in parallel and ignoring each other.
Tanishq Singh·May 27Freshers, the bar went up. The jobs didn’t go away.I get why freshers are panicking. I want to give you a different take.
Tanishq Singh·Jan 31A Thought on Constrained Thinking and AIA reflection on how we frame problems, not how machines solve them.
Tanishq Singh·Sep 23, 2025The Reality Behind Multi-Agent Systems: Why the Promise Doesn’t Match the PracticeMulti-agent systems are getting a lot of traction nowadays because of the promising nature of the work that they do. The promise is to…A response icon1A response icon1
Tanishq Singh·May 6, 2025From Documents to Datasets: Mastering Meta’s Synthetic Data KitHello Hello Hello!
