Aspiring Product Manager with 4.5+ years across B2B SaaS support, technical operations, and customer-facing roles.
I don't just write PRDs, I ship the product. This GitHub is proof of work: live products I've built end-to-end using AI-assisted and vibe coding tools, alongside the PRDs and case studies behind them.
- Turn support and customer signal into shipped product decisions
- Build and ship full products using vibe coding platforms (Lovable, Replit, Base44) and AI-assisted development (Next.js, Supabase, Claude API)
- Understand how engineering teams work, having partnered directly with engineers on requirements, scoping, and delivery
- Write PRDs using Working Backwards, JTBD, RICE, MoSCoW, Kano, and MECE frameworks
4.5+ years in B2B SaaS, moving progressively closer to the product itself at each step:
- eClerx, Analyst (2021 to 2023): process and data operations for enterprise clients, high-volume analyst workflows.
- Maropost, Client Support Specialist (2023 to 2024): owned discovery of a feature request workflow (JetSend) end to end through Google Analytics, analyzed 80+ tickets using MECE and JTBD interviews, wrote the requirements brief, shipped in 6 weeks, cited by 2 enterprise clients at renewal.
- Vartopia, Technical Support Engineer (2024 to 2025): resolved technical escalations for SaaS clients, working directly with engineering and account teams to close issues within SLA.
- HighLevel, Customer Support Representative, Remote US Market (Apr 2025 to Feb 2026): built and shipped an internal AI tool that cut Average Handling Time from 40 to 25 minutes and raised CSAT by 5%, and a separate AI support agent for white-label clients that cut ticket volume by 30% and raised CSAT by 20%, shipped in a 6-week cycle.
This is the background behind the PRDs and case studies below: I've sat close enough to the ticket queue and the escalation call to know what breaks for users, and close enough to engineering to know how to scope a fix that actually ships.
| Project | What It Is | Links |
|---|---|---|
| Loopback | B2B SaaS Voice of Customer tool. Ingests support tickets via CSV/API, runs them through the Claude API for synthesis, and delivers AI-generated weekly digests via Slack and email (Resend). Full stack: Next.js frontend, Supabase (Postgres) for auth/data, Vercel Cron for scheduled digest jobs, a provider-agnostic billing abstraction layer (Starter $49/mo, Pro $99/mo tiers), and a dashboard for usage and billing. Auth, onboarding, CSV upload, and email delivery all built end-to-end. | Live Product · Repo · PRD |
| RouCN Storefront | Full D2C sneaker brand storefront: product drops, sizing, cart, waitlist, brand story | Live Site |
| Portfolio | Case studies, certifications, and PM framework write-ups | Live Site |
| GoHighLevel Product Teardown | Competitive moat analysis and feature audit of a real SaaS product, written from 12 months inside the platform as a Technical Support Engineer | Live Site |
| Maropost Marketing Cloud Teardown | 9-section product teardown written from experience as Client Support Specialist, benchmarked against G2 and competitor data | Live Site |
| Project | What It Is |
|---|---|
| HighLevel AI Agent | AI agent for white-label clients: 30% ticket drop, 20% CSAT increase, shipped in 6 weeks |
| Maropost JetSend | Owned discovery through GA on a feature request workflow: 80+ tickets analyzed via MECE, JTBD interviews, requirements brief, shipped in 6 weeks, cited by 2 enterprise clients at renewal |
AI & Vibe Coding
Development
Product Management & Analytics
Collaboration & Docs
Frameworks
Support/Ops background: Ticket workflows, escalation handling, SLA management, cross-functional coordination with engineering
MBA, International Business Management, Amity University (2026)
LinkedIn · singhshiwal@gmail.com
Currently: Building Loopback and RouCN, actively looking for APM and PM roles.