👋 Hi, I'm Yash Dewasthale

Founder & Builder | 23 yo AI Software Developer from India Building tools that make developers more productive.

Currently focused on AI-powered developer tools and open-source infrastructure, and raising funds. I've also built LaminarFlow-an ai native intelligent layer.

🛠️ Skill

• Languages: TypeScript, JavaScript, Rust, Python, GO
• Frontend: Next.js, React, Tailwind CSS
• Backend: Node.js, PostgreSQL, Prisma
• Tools: Docker, Git, Turborepo, Pinecone, LangChain
• AI Research: FineTuning

🚀 Recent Products and Work

Supercode

CLI AI Coding Agent for developers

Work with superCli directly in your codebase. Build, debug, and ship from your terminal, IDE, Slack, or the web. Describe what you need, and superCli handles the rest.

Next · TypeScript · PostgreSQL · Pinecone · ThinkingMachine · HuggingFace · Fine-Tuning · Better Auth

LaminarFlow

AI-native Open-source Fintech Platform A modern, powerful, and affordable platform to manage your business' financial workflow.

Next.js · TypeScript · Prisma · PostgreSQL · AI-powered · Mastra · Elevenlabs

Devdocs

AI-native Open-source Documentation Platform Build beautiful, fast, and modern documentation sites with ease. MDX-powered with a powerful CLI.

Next.js 14 · MDX · Contentlayer · Turborepo · Radix UI

AI Music Generator

|AI-powered Music Generation Platform Generate music using AI. Create unique tracks with artificial intelligence.

Next.js · TypeScript · AI · Modal · HuggingFace · Fine-Tuning · Audio Processing

📦 Other Notable Work

• sourced-flow – Team management platform
• DocumentAndDraw – Developer tools for design and scale
• awesome-react-enaqx – React resources collection (19 stars)
• Web3-Developer-Path – Complete blockchain developer roadmap
• one1Letter – Marketing tool for developers

📝 Latest Blog Posts

• How Increasing Input Tokens Impacts LLM Performance - An in-depth exploration of how larger context windows affect LLM reasoning, latency, and cost. Learn the trade-offs between token limits and model performance.

• I Stopped Using Opencode and CodeRabbit for Real-World Dev Work - A candid take on why the author moved away from AI coding assistants for production development, and what tools they use instead.

• vLLM: The LLM Engine That Actually Treats GPUs With Respect - Deep dive into vLLM's PagedAttention mechanism and how it achieves groundbreaking throughput for LLM inference while optimizing GPU memory.

• Agent Skills: The Missing Layer Between Raw Models and Real Work - Exploring how agent skills bridge the gap between base LLM capabilities and practical developer workflows in AI-powered tools.

• Everything I Installed on My New Mac (2026) - A comprehensive setup guide covering the essential developer tools, apps, and configurations for a fresh Mac setup in 2026.

• How to protect your app from real work attacks - Learn how to secure your Next.js applications against common vulnerabilities using ArcJet - rate limiting, bot detection, and attack prevention made simple.

📈 GitHub Stats

📫 Connect

🧠 What I'm Doing

• Building AI-powered developer tools at Supercode
• Growing LaminarFlow as an open-source fintech solution
• Contributing to open-source through Devdocs
• Learning and exploring new AI/ML technologies

💰 Support

Building the future, one commit at a time.