Daily Dose of Data Science

Daily Dose of Data Science

Build a Jev Judge
Judgement does not need text generation...
11 hrs ago • Akshay Pachaar
MoE inference engineering, clearly explained
...with visuals.
Sep 23 • Avi Chawla
Build your own Jev (100% local)
...explained step-by-step with code.
Sep 22 • Avi Chawla
Jev, Clearly Explained
...with visuals.
Sep 21 • Avi Chawla
Run Any Agent Harness Under One Interface
How UHP and HarnessRouter standardize agent execution across Codex, Claude Code, Hermes, and other runtimes.
Sep 18 • Avi Chawla
Where does all the VRAM go during LLM inference?
...explained visually
Sep 17 • Avi Chawla
How to Fine-Tune LLMs in 2026
Reward-free RL is here!
Sep 16 • Avi Chawla
Easiest Way to Run Agent Harnesses Using Local Models
...explained with a full video walkthrough.
Sep 15 • Avi Chawla
What It Takes to Build a Production Agent Harness
A hands-on nanodegree for production agent engineering.
Sep 14 • Avi Chawla
4 Speculative Decoding Variants
...covered with tradeoffs in production.
Sep 12 • Avi Chawla
The Architecture for Serving 100 Fine-Tuned Models on One GPU
...explained with code.
Sep 11 • Avi Chawla
Why Multi-turn Agents Need More Than a Task Graph
What production agents must preserve between turns, and what they must reset.
Sep 10 • Avi Chawla
© 2026 Avi Chawla · Privacy ∙ Terms ∙ Collection notice
Start your SubstackGet the app
Substack is the home for great culture