I build systems
I can explain.
I'm Vatsal — a software engineer exploring the seam between fundamentals and modern tooling. From embedded alerts to grounded AI, I like projects that make the invisible visible.
Projects with receipts.
Six experiments across hardware, web, AI, and the beautifully low-level bits in between.
Smart Helmet — IoT Emergency Accident Alert System
A complete IoT-to-cloud workflow that detects abnormal acceleration, gives riders a cancellation window, and sends emergency email alerts with live location.
NanoNet — Neural Network From Scratch
A fully functional multi-layer neural network framework built entirely from scratch using only Python and NumPy.
Grounded RAG System for Technical Knowledge
A retrieval-augmented generation pipeline using vector embeddings, Qdrant, and query rewriting to prioritize grounded technical answers.
Death's Job — 2D Game in Pygame
A close-to-the-code arcade game with a hand-rolled loop, physics updates, collision handling, and state management.
Blog App — Next.js 14 Full-Stack Platform
A hybrid-rendered blogging platform with Server Components, MongoDB, NextAuth Google OAuth, pagination, and performance-minded queries.
Book Notes WebApp
A full-stack notes app for adding, editing, categorizing, searching, and sorting reading notes with Google OAuth and local auth.
Experience & Internships.
Hands-on software development, real customer impact, and production web engineering.
Software Engineer Intern
InternshipImproved the company’s website UI by refining layouts and enhancing user experience across key visitor flows.
Maintained and updated website content, ensuring continuous functionality, zero layout shift, and consistency across pages.
Fixed minor UI issues and collaborated actively with the team to implement production design improvements.
Useful at different layers.
A toolkit honed from silicon registers up to modern distributed cloud systems.
Things I'm thinking through.
Essays, architectural breakdowns, and field observations on systems and AI.
Why retrieval quality beats model size in technical RAG
A field guide to query rewriting, vector search, and making AI answers earn their confidence.
Shipping an IoT alert from sensor to inbox
What changes when your backend depends on a tiny board, a moving helmet, and an unreliable network.
The small joy of writing a neural net from scratch
Gradients stop feeling magical when you follow them one scalar at a time.
How I keep the signal clear.
Guiding tenets that preserve focus across messy domains and complex stacks.
Fundamentals First
Frameworks change quickly. Core concepts like algorithms, memory, and data flow do not. I focus on understanding the fundamentals so tools become interchangeable.
Abstractions Leak
Frameworks are useful, but understanding what's happening underneath makes debugging, optimization, and design decisions much easier.
Active Construction
Learners are not empty vessels; they build new understanding by connecting prior experience with fresh information.
Have a good
problem?
No pitch decks required. Just a clear question.