Writing

Multi-Task RL Training: Consolidating Specialized Agents into One Model
Can one RL-tuned model master very different tasks without forgetting? Our ART-E and Summary-RL experiments show minimal forgetting and strong interleaved performance.
27 Aug 2025
Render Mermaid Diagrams in Astro
Adding Mermaid Diagrams support to Astro
08 Apr 2025
From RAG to Agents: The RL moment
How simple rewards could transform retrieval systems into emergent search agents
02 Feb 2025
Scaling Compute for RAG
Unlock higher RAG accuracy by strategically spending compute.
19 Jan 2025
Render LaTeX in Astro
Adding LaTeX support to Astro
14 Dec 2024
My Git Rebase Workflow
23 Sept 2021
End-to-End Object Detection with Transformers
29 May 2020
Splitting comma seperated fields into multiple rows in SQLite
03 Nov 2019
Deep Contextualized Word Representations
15 Feb 2019
Monocular Dense Reconstruction
27 Apr 2018
JavaScript WebView and iOS
20 Mar 2018
Using Git for Word
27 Dec 2017
Hosting a Django application on AWS for free
15 Dec 2017
How to achieve 100% test coverage in React
30 Sept 2017
Bringing Technology to Construction
11 Aug 2017
Hawker Centres in Singapore
05 Apr 2017
Debugging in python using print statements
12 Mar 2017