

I BUILD SYSTEMS
THAT OUTLIVE TRENDS
I approach engineering as a long-term craft — where clarity, performance, and scalability are not features, but foundations.
My work focuses on building systems that remain stable under pressure, adaptable to change, and intuitive for users.
I prioritize the product, the experience, and the problem — placing them above myself.
The code is temporary. The system is not.
About / Identity
01Full Stack Engineer
02Builder of scalable systems
03Focused on performance & UX
SELECTED WORKS
01
RoleCrest
AI HIRING & TALENT INTELLIGENCE
02
Lyzr AI Platform Revamp
ENTERPRISE AI & 3D WEB ARCHITECTURE
01
RoleCrest
02
Lyzr AI Platform Revamp




RoleCrest
[ THE ARCHITECTURE ]
Architecting an end-to-end recruitment intelligence platform that eliminates resume-screening friction through automated talent scoring, direct recruiter synchronization, and real-time interview evaluation.
Engineered an AI-native hiring ecosystem featuring 1-click LinkedIn/Google OAuth profile enrichment, a multimodal resume parsing pipeline (PDF/DOCX/image), and real-time interactive technical mock interviews with automated candidate fitness scoring. Designed a high-throughput recruiter management interface with transparent salary bands and live application tracking, replacing passive job boards with direct, verified candidate-to-employer pipelines.




Lyzr AI Platform Revamp
[ THE MISSION ]
Architecting a high-performance, immersive 3D web platform for enterprise AI agent productionization, governance, and real-time observability.
Engineered an enterprise-grade AI agent productionization platform featuring an interactive 3D layer visualizer, dynamic telemetry HUDs, split-matrix governance comparisons, and fluid GSAP/Lenis animations. Maintained ultra-fast page performance with a 352ms First Paint, 26ms TTFB, and 946ms complete load time, delivering silky 60 FPS motion across 3,000+ DOM nodes and complex 3D scenes.
LATEST WRITINGS
Blogs
& Writings
If you’re not obsessing over LCP, FID, and CLS — you're leaving users (and revenue) on the table.
5 min readDebounced Search with Client-side Filtering: A Lightweight Optimization for Large Lists
If you’re not obsessing over LCP, FID, and CLS — you're leaving users (and revenue) on the table.
5 min readDebounced Search with Client-side Filtering: A Lightweight Optimization for Large Lists
PerformanceIf you’re not obsessing over LCP, FID, and CLS — you're leaving users (and revenue) on the table.
Why Speed Metrics Matter (Beyond Just "Feeling Fast") Performance isn’t just "loading fast." FCP, LCP, FID, CLS : These aren't just buzzwords; Core Web Vitals are backed by real user studies. It’s about how users experience your site in those crucial first few seconds.
PerformanceDebounced Search with Client-side Filtering: A Lightweight Optimization for Large Lists
Recently ran into a case where I had to support 1500+ cities in an autocomplete. Rendering all at once was janky. Virtualization felt overkill. So I optimized it with a simple debounced client-side search — no extra complexity, just lodash.debounce + in-memory filtering. > Smooth UX > Lightweight > Easy to reuse across search inputs
SKILL EXPERTISE
Skills
& Expertise
Technical
Arsenal
Building robust, scalable systems and immersive digital experiences. A curated stack focused on performance, aesthetics, and modern engineering practices.
Languages
Frameworks & Databases
Cloud & DevOps
Libraries & Tools
AI Tools
GET IN TOUCH
Let's start a
project together
DIRECT INQUIRY
For project kickoffs, fast discussions, or time-sensitive inquiries:
WhatsApp Messenger
DIRECT INQUIRY
For project kickoffs, fast discussions, or time-sensitive inquiries:
WhatsApp Messenger