Portfolio · 3 featured projects

Hi, I'm Nishchal. I build software for the web.

Backend Engineer · AI Systems · Full-Stack Developer. A selection of projects I've shipped solo or as part of a team — case studies, code, and the occasional hard-earned lesson.

Explore projects

10+

Freelance projects shipped

1000+

DSA problems solved

9.43

CGPA · CSE (AI & ML)

v.2026

// currently shipping

Multi-role CRM platforms
RAG + voice AI pipelines
Production Next.js builds for clients

Latest project

GDF Internationals / Bee CRM · GGEZZ · Real-Time AI Voice Agent

Featured work

Selected projects.

A handful I'm proudest of. Click through for the full case study with context, decisions, and outcomes.

CRM2026

/ 01 · Lead engineer (solo build)

GDF Internationals / Bee CRM

High-velocity CRM built for concurrent dialers on resource-constrained hosting. MongoDB L2 snapshotting, per-lead save queues, and visibility-aware polling keep a 40K+ lead pool responsive across Admin, Employee, Advisor, and Case Assessor roles.

DashboardStats snapshotting — 24+ aggregations replaced by one document read (~5.5s → 100–300ms)
Delta-based board sync (`?since=`) plus per-lead serial save queue — no disposition write collisions
Visibility-aware polling pauses background tabs; offset pagination guard (50/page) avoids expensive counts
Full floor lifecycle: bulk lead ingest, CRM OTP gating, advisor escalation, compliance assessors, commission payroll

Next.js 16React 19TypeScriptPrisma 6MongoDBjose (JWT)CloudinaryTailwind CSS v4Framer Motion

Read case study

Confidential deploymentLive link withheld — production CRM running for GDF Internationals. Code, architecture, and walkthroughs available on request.

AI2026

/ 02 · Solo build

Real-Time AI Voice Agent

Express backend that closes the loop from spoken intent to grounded natural language, then back to synthesized speech — keeping the same RAG stack usable from plain text clients.

Layered async pipeline: STT → embed → vector search → LLM → TTS, with bounded retries and graceful degradation
PDF ingestion + 1000+ embeddings (sentence-transformers / 384-dim Pinecone), ~40% better retrieval accuracy
Multi-tenant by design — retrieval is namespace-filtered by `avatarId` so knowledge stays isolated
Fail-soft policy: TTS down → text still returns; RAG down → safe message instead of hallucination

Node.jsExpressDeepgramPineconeGroqHugging FaceMongoDBJWT

Read case study

Not deployed publiclyBackend reference implementation — currently runs locally / on private infra. Architecture deep-dive and code walkthrough available on request.

Full Stack2026

/ 03 · Solo full-stack engineer

GGEZZ — Premium Gaming Cafe

Production-grade management suite for PC bangs and console arenas. Minute-accurate session timing, food orders merged into one checkout invoice, staff floor panels, and a cyberpunk-themed player portal with play history and stats.

Consolidated billing — console playtime + cafe items on one itemized checkout invoice
Millisecond-accurate session timing at Rs 100/hour; hidden iframe print flow with no navigation interrupt
Reactive provider pattern keeps staff dashboards live; role gates for `(staff)` vs `(player)` route groups
Peak-hour analytics, top-spending members, and automated welcome + checkout emails via SMTP

Next.js 16React 19TypeScriptMongoDBTailwind CSS v4NodemailerPDFKit

Read case study Live site

View all projects

Experience

Where I've shipped.

Full-time, freelance, and the in-between.

Freelance Full-Stack & AI Engineer

Self-employed · Remote · India

2024 — Present

Partnered with 10+ clients across EVs, architecture, cinema, NGO, consulting, and community organizations to ship production Next.js platforms.
Designed and built RAG / voice AI systems, multi-role CRMs, payment-integrated marketplaces, and content-driven sites end-to-end.
Average engagement: 3–6 weeks. Solo build, full handoff with documentation.

Next.jsTypeScriptMongoDBRazorpayCloudinaryPineconeGroqDeepgram

Virtual Intern — Springboard 6.0

Infosys · Virtual

Jan 2026

Completed 20+ enterprise modules covering system design, backend workflows, and scalable application development.
Applied real-world concepts in projects, deepening understanding of enterprise architecture and delivery practices.
Collaborated in structured environments following industry-standard delivery and documentation practices.

System DesignEnterprise ITBackend WorkflowsDocumentation

Frontend Developer

Tech Instance · India

Dec 2024 — Jan 2025

Architected an Electron.js desktop platform that improved user retention by 35% via optimized UX workflows.
Implemented secure checkout systems that lifted transaction success rates by 40%.
Optimised UI performance and load handling, cutting page load time by 25%.

Electron.jsReactJavaScriptE-commerce

Education

Where I learned the craft.

Sep 2024 — Jun 2028

B.Tech — Computer Science & Engineering (AI & ML)

Ajay Kumar Garg Engineering College (AKGEC), Ghaziabad

Result: SGPA 9.59 / Branch Topper (9.4 CGPA)

Specialisation in AI & ML; strong systems & backend coursework foundation.
Solved 1000+ DSA problems — Global GFG Rank 99, active on LeetCode & CodeStudio.
Active GirlScript Summer of Code (GSSoC) 2025 contributor — backend track.
Finalist at AI India Innovates Hackathon 2026.

Apr 2019 — Mar 2023

Higher Secondary Education (Class 12) — PCM

Durgaprasad Vidyaniketan

Result: 94%

PCM stream with 94% overall.
Foundation for engineering and competitive problem-solving.

Get in touch

Have a project in mind, or want to chat code?

I'm always up for a good conversation about software, startups, or where to find the best coffee in your city. Send a note and I'll reply soon.

sachannishchal@gmail.com

Location

Ghaziabad, Uttar Pradesh, India · Remote-friendly