$ user.role_

Abhinav Shyju _

> Software Engineer

$ user.description_

const profile = {
role : "Full Stack & AI Engineer",
focus : "Full-stack development, AI/ML, and scalable customer-facing products",
expertise : ["full-stack", "ai/ml", "frontend"],
experience : "2+ years"
}

$ cat user.summary_

# output: Software Engineer with experience building full-stack applications and AI-driven customer experience platforms. Worked across React, TypeScript, PostgreSQL, Cloudflare, and modern AI stacks including RAG pipelines, vector databases, conversational memory systems, and multi-model LLM integrations.|

$ user.socials_

$ scroll to explore

Professional Journey

A timeline of my career progression

career_timeline.sh

~/career cat career_timeline.txt
2025 - Present Founding Software Engineer (AI & Full Stack) @Pure Intelligence AI Technologies
$ As founding engineer at Pure Intelligence AI Technologies, I built MagicalCX from scratch — an empathy-first AI customer support platform that lets businesses handle support like a great human would, but faster and across every channel. Shipped context-aware AI agents with conversational memory, a unified omnichannel inbox (web, email, WhatsApp, Instagram, Messenger), seamless human handoff with full context, the EFRO™ upsell engine, and reporting dashboards.
# # Technologies used:
Next.js
TypeScript
Firebase
Qdrant
VAPI
OpenAI
Claude
Gemini
2024 - 2025 Full Stack Developer Intern @InvestorsMate
$ Contributed to a property investment platform serving 2,000+ active users. Built frontend features with Astro, React, and Shadcn UI; developed backend services with Cloudflare Workers and PostgreSQL; and participated in database schema design and query optimization.
# # Technologies used:
Astro
React
TypeScript
Drizzle
Cloudflare
PostgreSQL

Projects

A gallery of my projects, showcasing my skills and creativity.

MagicalCX

Empathy-first AI customer support — omnichannel inbox, memory, handoff, and EFRO™ upsells.

As founding engineer, I built MagicalCX from scratch — an empathy-first AI customer support platform with human-like agents, conversational memory, a unified omnichannel inbox, seamless human handoff, the EFRO™ upsell engine, and reporting dashboards on a Next.js stack with Firebase, Qdrant, and multi-model LLM tooling.

Next.js TypeScript React Firebase Qdrant VAPI +5 more

Loofy AI

In progress

Meeting-bot SaaS — Google Meet today, Zoom and Teams next. Live transcription, speaker ID, bidirectional audio, and API-driven post-meeting workflows.

I built Loofy AI as a meeting-bot SaaS — Google Meet today, with Zoom and Teams planned. The bot captures live transcription and speaker ID, exposes post-meeting workflows via API and dashboard, supports bidirectional WebSocket audio for in-meeting AI, and runs multi-model orchestration (LLMs, STT, TTS) on a FastAPI + PostgreSQL stack.

FastAPI PostgreSQL Python WebSockets Vite React +4 more

OrdraDesk

Web order desk for distribution companies — role-based workflows across field reps, office staff, and distributors, with AI-assisted order management.

I built OrdraDesk as a web order desk for distribution companies — field reps capture shop orders, office staff arrange and assign them to distributors, and role-based access replaces paper workflows. Next.js, Firebase, better-auth, TanStack Query, and AI SDK (Groq) with Excel/PDF exports.

Next.js TypeScript React Firebase better-auth TanStack Query +3 more

TeleCRM

Pocket call center for Techgebra — one shared number, 100+ callers, no telephony hardware.

Built TeleCRM for Techgebra — a pocket call center that replaces traditional telecaller hardware. One shared number, 100+ mobile callers, lead upload, outbound dialing, and inbound callbacks from a three-app monorepo (Expo, Express, React admin) with LiveKit + Plivo PSTN.

Expo React Native Express TypeScript PostgreSQL Drizzle ORM +4 more

cursor-openai

Drop-in OpenAI-shaped client for Cursor models — LangChain and LangGraph compatible.

Drop-in OpenAI-shaped client for Cursor models. Pass a Cursor API key + model id; call chat.completions.create or use LangChain / LangGraph. In-process via @cursor/sdk — no Express gateway.

TypeScript @cursor/sdk LangChain LangGraph Node.js
View all projects →

contact.sh

~/contact cat contact.txt

# Let's Connect!

Currently building AI-powered customer experience platforms at Pure Intelligence AI. Open to discussing interesting projects and opportunities.