Headline

Product Analyst | SQL · Python (pandas) · Metabase · Mixpanel | Built UrbanIQ (4K+ data points) | CS Engineer · GenAI | Open to PA / Growth Analyst Intern

About

I build products and then analyse what actually happens when real users touch them. Final-year CS student at IKGPTU. I shipped UrbanIQ — a complaint and document management platform deployed at my university, processing 4,000+ submissions. Then I ran EDA on that data (Python, pandas, Metabase) to find where the system was breaking down: SLA breaches by department, resolution time bottlenecks, category spikes. That analysis fed back into product decisions. Currently building DreamAI — an AI image generation platform — and designing its full analytics stack from scratch: Mixpanel event instrumentation, D7 retention cohorts, funnel drop-off analysis, onboarding flow iteration. My stack as a PA: → SQL (JOINs, CTEs, window functions) → Python: pandas, matplotlib, numpy, Jupyter → BI: Metabase, Google Looker Studio → Product analytics: Mixpanel, funnel analysis, cohort analysis, AARRR → Technical depth: React, Next.js, TypeScript, GenAI integration I care about one thing: turning messy data into a decision that makes the product better. Open to Product Analyst internships and Growth Analyst roles across India. Portfolio: shivam-sharma-portfolio-theta.vercel.app

Location

Ludhiana

AI stack

cursor

hugging face

openai

replit

Work experience

Founder & Product Engineer at Cute Helper, Jan 2026 - Present · 6 mos, Remote, Transparent, always-on-top desktop AI companion — built so AI teaches you, not replaces you. What it does: → Floating animated mascot lives on your screen, invisible except for the character itself → Click once → speak → it transcribes (Groq Whisper), optionally reads your screen, responds via glassmorphism panel with TTS → Maintains full conversation context across a session; auto-listens for follow-ups → Never interrupts workflow — passes clicks through transparent areas to whatever's behind it Analytics layer (in progress): → Tracking avg. query length, follow-up rate per session, silence detection accuracy, and response gap metrics → Goal: identify where the "teach don't do" loop breaks down in real usage Stack: Electron · React · TypeScript · Vite · Groq Whisper · LLaMA 4 Scout (vision) · Web Speech Synthesis · Web Audio API · desktopCapturer

Product Manager – Independent Projects at Independent Projects, Aug 2025 - May 2026 · 10 mos, Maharashtra, India, Leading product strategy and execution across two domains — civic tech (UrbanIQ) and gaming (Valorant, Genshin Impact). Conducted user research, opportunity sizing, competitive analysis, and feature prioritization. Published two in-depth PM case studies demonstrating end-to-end product thinking from problem discovery to solution design.

Founder & Solo Developer at UrbanIQ, Jul 2025 - May 2026 · 11 mos, Ludhiana, Punjab, India, Built a full-stack AI-powered complaint and document management platform designed for institutional use — complaint routing, document approvals, role-based dashboards, and AI-assisted complaint categorisation. → Designed and shipped end-to-end solo: PRD → architecture → frontend → backend → deployment → Built AI-driven complaint classification to auto-route submissions to the correct department → Authored 15+ user stories across 3 stakeholder roles (student, staff, admin); maintained sprint velocity using Agile → Reduced simulated resolution time by 30% in load testing through workflow optimisation Stack: Next.js · React · TypeScript · Supabase · Prisma · REST APIs · Jira · Notion

Education

Bachelor of Technology - BTech, Computer Science, I. K. Gujral Punjab Technical University (IKGPTU)

machine learning certification course, Machine Learning, Internshala Trainings