I build the parts of a product that are hard to fake: real-time systems that hold up under load, frontend architecture that scales past the demo, and data pipelines tuned down to the millisecond. 3+ years shipping to 10,000+ daily active users, from WebRTC-powered collaboration tools to LLM-integrated platforms.
A little about how I work
The short version: I go looking for the bottleneck nobody assigned me to fix, then ship a measurable improvement — not just a feature.
Hi, I'm Chintan. I'm a software engineer with 3+ years building scalable, high-performance web applications — most recently as a Project Supervisor guiding teams through cloud-native SaaS and LLM/RAG integrations.
My depth is in the systems most engineers avoid: real-time communication (WebRTC, MQTT), binary data optimization (Protocol Buffers, FlatBuffers), and design systems that hold up at scale. At Treeleaf Technologies I helped take anydone to 10,000+ daily active users, cutting real-time latency by 15% and data transmission latency by 40%.
I also write about what I build — my article on FlatBuffers performance in JavaScript is linked below.
Selected work
Real-time systems, AI-integrated tools, and frontend platforms shipped to production — not tutorial clones.
Interactive real-time collaborative whiteboard built with HTML Canvas and Web Workers, syncing state across up to 50 concurrent users with binary-efficient payloads.
Core real-time communication layer for an AI-integrated workflow platform, built on Janus WebRTC SFU with MQTT + Protobuf transport — 40% more stable, 15% lower latency.
Full-stack app that turns natural-language descriptions into production-ready UI, using GPT-4 to generate semantic HTML with ARIA accessibility and BEM conventions baked in.
Privacy-first image background remover running entirely in-browser via Transformers.js, with optional WebGPU acceleration for RMBG-1.4 and MODNet models.
Atomic-design search portal aggregating product listings across Nepalese e-commerce platforms (Daraz, HamroBazaar & more), with collaborative-filtering recommendations.
Workforce analytics platform for productivity tracking — built key dashboard components with Redux Saga/Thunk, improving load times by 15%.
Where I've worked
Most recent first — like any good log.
Guided teams architecting cloud-native SaaS and advanced AI features (LLM integration, RAG pipelines). Mentorship reduced technical debt by ~10% and drove adoption of secure, high-performance design patterns.
Engineered core communication services for 'anydone' via Janus WebRTC SFU — 40% more stable real-time connections, 15% lower latency for 10,000+ daily active users. Built the collaborative canvas board and integrated MQTT + Protobuf/FlatBuffers, cutting data transmission latency 40%. Design systems work lifted user satisfaction 15% and engagement 20%.
Optimized Node.js/Express backend services, improving API response times by 15%. Contributed to Java Spring Boot features and bug fixes, reducing reported issues by 10% through rigorous code review.
Built and optimized ReactJS components with Ant Design and Tailwind CSS, improving page load times by 10%. Streamlined form UX with Formik, cutting submission errors by 20%.
Education & writing
Formal training plus one thing I published because I couldn't stop thinking about it.
Master of Information Communication Technology — University of the Sunshine Coast
Artificial Intelligence, Data Science, System Design, Cybersecurity, Databases.
B.Sc (Hons) Computing, First Class Honours (80.39%) — London Metropolitan University
Software Engineering, Data Structures & Algorithms, Advanced Databases, Operating Systems.
Maximizing Performance with FlatBuffers in your JavaScript Application
A deep dive into using FlatBuffers to cut serialization overhead and data transmission latency in real-world JS apps — the same technique behind the 40% latency reduction on 'anydone'.
Let's build something
Have a project in mind, or just want to say hi? My inbox is open.