import { useState } from "react"; import { AnimatePresence, motion } from "framer-motion"; import { ExternalLink, FileText, Github, Linkedin } from "lucide-react"; import { CONTACT_EMAIL, GITHUB_URL, HUGGING_FACE_URL, LINKEDIN_URL } from "@/lib/links"; type Project = { id: string; era: "UNT - Graduate" | "Accenture" | "Undergrad" | "Independent"; title: string; tagline: string; period: string; details: string[]; stack: string[]; links?: { label: string; href: string; }[]; }; const projects: Project[] = [ { id: "alive-agents", era: "Independent", title: "Alive - AI Agent Orchestrator", tagline: "Full-stack platform that assembles AI dev teams for any coding task.", period: "2026", details: [ "Spawns 5 role-based agents (Team Lead, Architect, FE Dev, BE Dev, Tester) per project.", "Each agent runs on a different AI provider (DeepSeek, Qwen, Kimi, GLM, GPT) - all in parallel.", "Isolated git worktrees per agent, real-time progress tracking, and live iframe preview.", ], stack: ["TypeScript", "Next.js", "Multi-Agent AI", "React", "Prisma"], links: [ { label: "View Code", href: "https://github.com/im-mj/Agents-", }, ], }, { id: "nimbus-bank", era: "Independent", title: "Nimbus Bank - AI Triage System", tagline: "LangGraph multi-agent pipeline for banking customer support.", period: "Apr 2026", details: [ "Built a LangGraph multi-agent system to classify and route banking support tickets.", "Agents handle intent detection, policy lookup, escalation, and resolution - end to end.", "Delivered full PRD, architecture diagram, and live Streamlit demo for Wipro pre-screening.", ], stack: ["Python", "LangGraph", "LLM", "Streamlit", "Multi-Agent"], links: [ { label: "View Code", href: "https://github.com/im-mj/Nimbus-Bank-Triage", }, ], }, { id: "ghost-job-hunter", era: "Independent", title: "Ghost Job Hunter", tagline: "Chrome extension concept for spotting suspicious or stale job posts.", period: "2026", details: [ "Flags ghost-job warning signs such as stale listings, repeated postings, vague role signals, and weak hiring intent.", "Uses an AI-assisted scoring flow to help job seekers focus on real opportunities instead of wasting applications.", "Connects analytics, recruiting pain points, and browser automation into a practical job-search tool.", ], stack: ["Chrome Extension", "JavaScript", "AI Scoring", "Job Analytics"], links: [ { label: "View Code", href: "https://github.com/im-mj/Ghost-Job-Hunter", }, ], }, { id: "smartscreen", era: "UNT - Graduate", title: "SmartScreen - AI Resume Analyzer", tagline: "AI-driven resume scoring & ATS-match feedback.", period: "Jan 2025 - May 2025", details: [ "Designed an AI scoring engine that grades resumes against a target JD and highlights ATS-killing gaps.", "Generates targeted feedback so candidates can iterate quickly instead of guessing.", "Built as an end-to-end concept: parsing -> scoring -> recommendations.", ], stack: ["Python", "NLP", "AI", "Data Analysis"], }, { id: "cvs", era: "UNT - Graduate", title: "CVS Pharmacy Strategy Thesis", tagline: "Analytics-backed strategy across Finance, Marketing, Ops, HR & AI.", period: "Jan 2025 - May 2025", details: [ "Deep-dived 10-K filings and market data to build a cross-functional growth thesis.", "Layered AI/automation recommendations on top of traditional strategy frames.", "Delivered as a board-style presentation with prioritized initiatives.", ], stack: ["Research", "10-K Analysis", "Strategy", "Presentations"], }, { id: "jobs", era: "UNT - Graduate", title: "Job Market Analysis", tagline: "Interactive Tableau dashboards on hiring trends & skills demand.", period: "Aug 2024 - Dec 2024", details: [ "Cleaned and modeled a multi-source job postings dataset.", "Built dashboards that surface skill demand, salary bands, and role trajectories.", "Designed for non-technical stakeholders to filter and self-serve.", ], stack: ["Tableau", "Data Viz", "SQL"], }, { id: "spend", era: "UNT - Graduate", title: "Student Spending Analysis", tagline: "Regression + ANOVA on monthly student spend patterns.", period: "Aug 2024 - Dec 2024", details: [ "Collected and cleaned a primary dataset from student respondents.", "Used regression and ANOVA to identify which factors actually moved spend.", "Translated stats into plain-English recommendations.", ], stack: ["Minitab", "Excel", "Statistics"], }, { id: "smile", era: "UNT - Graduate", title: "Smile Buddy 3000 - Marketing Concept", tagline: "Superhero-themed product campaign with positioning + messaging.", period: "Jul 2025 - Sep 2025", details: [ "Built persona, value prop, and a launch narrative around a fun mascot.", "Designed campaign assets and a phased go-to-market plan.", ], stack: ["Marketing Strategy", "Brand", "Advertising"], }, { id: "secondlife", era: "UNT - Graduate", title: "Second Life - Organ Matching ML", tagline: "ML pipeline for patient-organ matching and hospital triage.", period: "2025", details: [ "Built end-to-end ML pipeline with tiered matching, feature engineering, and model evaluation.", "Modeled patient and hospital journeys; benchmarked classifiers for match scoring accuracy.", "Deployed on HuggingFace Spaces with full architecture diagrams and a presentation deck.", ], stack: ["Python", "ML", "Flask", "SQL", "HuggingFace"], links: [ { label: "View Code", href: "https://github.com/im-mj/SecondLife", }, ], }, { id: "firewall-auto", era: "Accenture", title: "Firewall Rule Automation", tagline: "Cut manual firewall-rule work by ~40% across enterprise estates.", period: "2022 - 2024", details: [ "Wrote scripts to push and validate firewall rule updates across multiple devices.", "Reduced manual ticket cycle time and configuration drift.", "Improved rule consistency across Telstra Global's hybrid environments.", ], stack: ["Python", "Shell", "Firewalls", "Automation"], }, { id: "malicious-url", era: "Undergrad", title: "Malicious URL Detection (ML)", tagline: "Final-year ML project classifying URLs as safe vs malicious.", period: "Nov 2020 - Feb 2021", details: [ "Led a team during lockdown to build, train and evaluate the classifier.", "Engineered URL features and benchmarked multiple ML models.", "Presented results in multiple seminars to faculty and peers.", ], stack: ["Python", "Machine Learning", "Team Lead"], links: [ { label: "Research Details", href: "#research-publications", }, ], }, { id: "mask", era: "Undergrad", title: "COVID Face Mask Detection", tagline: "Computer vision classifier on a custom dataset.", period: "Mar 2020 - May 2020", details: [ "Built a face-mask detector with a friend-collected custom dataset.", "Iterated remotely during the early pandemic.", ], stack: ["Python", "ML", "Computer Vision"], }, { id: "eeg", era: "Undergrad", title: "Emotion Recognition from EEG Signals", tagline: "Internship concept project applying ML to EEG.", period: "Dec 2019", details: [ "Explored ML approaches for classifying emotions from EEG signals.", "Collaborated with multi-location team and a hackathon.", ], stack: ["Python", "ML", "Signal Processing"], }, ]; const researchDetails = [ "Presented at BIGS 2025 and published for SWDSI 2026.", "Built around AI-driven malicious URL detection using deep learning and BERT architecture.", "Connects cybersecurity experience with machine learning research and business analytics storytelling.", "Shows research writing, model evaluation, and technical presentation experience.", ]; function HuggingFaceMark() { return ( HF ); } const workLinks = [ { label: "LinkedIn", href: LINKEDIN_URL, note: "career profile", icon: , }, { label: "GitHub", href: GITHUB_URL, note: "code projects", icon: , }, { label: "Hugging Face", href: HUGGING_FACE_URL, note: "AI work", icon: , }, { label: "Research Paper", href: "#research-publications", note: "BIGS & SWDSI", icon: , }, ]; const eraColors: Record = { "UNT - Graduate": "bg-accent/15 text-accent border-accent/30", Accenture: "bg-chakra/15 text-chakra border-chakra/30", Undergrad: "bg-navy/10 text-navy border-navy/20", Independent: "bg-emerald-600/10 text-emerald-700 border-emerald-600/25", }; export function FeaturedProjectsSection() { const [open, setOpen] = useState(null); const [researchOpen, setResearchOpen] = useState(false); return (

PROJECTS

Featured Work

Hover or tap a card to read the story behind it.

{workLinks.map((link) => ( {link.icon} {link.label} {link.note} ))}
setResearchOpen((current) => !current)} onKeyDown={(event) => { if (event.key === "Enter" || event.key === " ") { event.preventDefault(); setResearchOpen((current) => !current); } }} onMouseEnter={() => setResearchOpen(true)} onMouseLeave={() => setResearchOpen(false)} layout className={`group mt-8 cursor-pointer rounded-2xl border bg-card px-6 py-5 shadow-sm transition-all ${ researchOpen ? "border-accent shadow-xl shadow-accent/10" : "border-accent/25 hover:border-accent/40" }`} >

Research Publications

AI-Driven Malicious URL Detection using Deep Learning and BERT Architecture.

Published at BIGS 2025 & SWDSI 2026
{researchOpen && (
    {researchDetails.map((detail) => (
  • {detail}
  • ))}
)}
{projects.map((p) => { const isOpen = open === p.id; const projectLinks = p.links ?? [ { label: "Request Details", href: `mailto:${CONTACT_EMAIL}?subject=${encodeURIComponent(`Project details: ${p.title}`)}`, }, ]; return ( setOpen(isOpen ? null : p.id)} onKeyDown={(event) => { if (event.key === "Enter" || event.key === " ") { event.preventDefault(); setOpen(isOpen ? null : p.id); } }} onMouseEnter={() => setOpen(p.id)} onMouseLeave={() => setOpen(null)} layout className={`group relative text-left rounded-2xl border bg-card p-5 overflow-hidden transition-all ${ isOpen ? "border-accent shadow-xl shadow-accent/10" : "border-border hover:border-accent/40" }`} >
{p.era} {p.period}

{p.title}

{p.tagline}

{isOpen && (
    {p.details.map((d) => (
  • {d}
  • ))}
)}
{p.stack.map((s) => ( {s} ))}
{isOpen ? "-" : "+"}
); })}
); }