index / 03 - blog archive

blog

archive stats
12
total
10
this yr
36K
words
ALLENGINEERINGSYSTEMSSECURITY
grep "..."
Two Gates and a Process Boundary: Swapping vLLM's Suffix Drafter Without a Fork
vLLM gates suffix decoding behind arctic_inference in two places, in two different processes. Replacing the drafter with my own meant clearing both - and getting there took one Python import-binding mistake and one process-boundary mistake.
engineering07/23/269 min
slate: Turning a Phone into a Stream Deck for macOS - One Command, a Pluggable Backend, and the Trade-offs Behind It
How I built slate - a free, open-source phone-as-Stream-Deck for Mac that runs over your LAN with no cloud. One semantic Command, a pluggable Transport, a Zod schema shared across three languages, and the trade-offs (including why not WSS) behind it.
engineering07/01/2620 min
No Public Ports: Zero-Trust Team SSH to a Private Azure Cluster with a Reverse Bastion and Cloudflare WARP
How I gave a whole team SSH to a private Azure dev cluster with zero internet-facing ports - one outbound-dialed cloudflared bastion, Cloudflare WARP, and Google SSO - plus the split-tunnel routing trap I hit on the way.
systems07/01/2617 min
Bear the Tokens: FP16 and Suffix Decoding to 9,000+ Tokens a Second on a Tesla T4
How I pushed Qwen2.5-0.5B to 9,052 tok/s on a free Colab Tesla T4 in the H2Loop inference contest - from AWQ INT4 dead ends to FP16 plus model-free suffix speculative decoding.
engineering06/14/2621 min
Logpipe: Building a Distributed Logging Stack from Scratch with Vector and ClickHouse
How I built Logpipe - a small but real Vector + ClickHouse logging pipeline in Go. Why batching at an aggregator is the whole point, what 'Too many parts' actually means, and how disk buffers stop you from losing logs on a restart.
engineering06/14/2625 min
TapTalk: Building a Free, Local, Open-Source Alternative to SuperWhisper and Whisper Flow
How I built TapTalk - a fully local, privacy-first speech-to-text app for macOS in Rust and SwiftUI. On-device Whisper, Core ML on the Neural Engine, chip-tuned inference, and a global push-to-talk hotkey. No cloud, no subscription.
engineering05/24/2621 min
The Three Levers: Backend-Driven UI, OTA Updates, and the End of Waiting for App Review
How Swiggy, Airbnb, PhonePe, and Nubank ship UI changes without waiting for app store approval - and how I built a working BDUI + OTA system in React Native with Cloudflare Workers and EAS.
engineering04/30/2618 min
How a Roblox Cheat Script Took Down Vercel: Anatomy of a Supply Chain Breach
Dissecting the April 2026 Vercel breach - from infostealer malware to OAuth hijack to customer data exposure. Full kill chain, trust lessons, and survival guide.
security04/21/269 min
Building a Real-Time F&O Fair Value Engine for Indian Markets
How we built a production fair-value calculator on top of Dhan - Black-Scholes, cost of carry, live depth, and mispricing signals - for Saras and NSE F&O.
engineering04/11/267 min
Building a World-Monitor News Engine for Traders: From ~115 Public Sources to a Daily Brief
How we built an agentic pipeline at Saras to scrape public macro and trade intelligence, dedupe at scale, enrich with structured AI, and ship a trader-facing brief - plus why Scrapling's StealthyFetcher sits in the hot path.
engineering04/11/267 min
Building a Memory Engine on Object Storage: A Deep Dive
Exploring how modern data systems separate storage and compute, with a hands-on implementation of a memory engine that works on object stores like S3.
systems12/29/2413 min
Reproducing CVE-2025-55182: Remote Code Execution in React Server Components
A deep dive into the critical RCE vulnerability affecting Next.js and React 19, with a full reproduction and technical breakdown.
security12/15/245 min
showing 12 of 12 posts
© 2026 Dhruv VakharwalaNext.js · TypeScript · Tailwind · PartyKit