Each one: the problem, what I built, what happened. Client names are kept private.
Hospitality · 2026Live sales dashboard for a new café
Problem: owners only saw sales at the end of the day, in a billing program they couldn't easily read.
Built: a small program that copies every bill into Notion every 5 minutes, with dashboards for daily sales, payment mix, busy hours and top dishes, plus a check for missing bill numbers.
Result: matched the billing software's own report to the rupee on day one. Owners now see the day as it happens.
NotionAutomationPowerShellDashboards
Events · 2026Music competition system, sign-up to results
Problem: a first-time competition with registrations, payments, judges and stage plans, run by a tiny team.
Built: a 40-question registration form, participant database, Razorpay payment tracking, automatic emails, judge score sheets, certificates, schedules and on-ground briefs.
Result: 21 registrations, a 4-day first round, and a second round with 17 acts, all tracked in one place.
NotionRazorpayApps ScriptEvent ops
Operations · 2026One Notion workspace for a whole company
Problem: tasks, menus, vendors, finance and staff updates lived in WhatsApp and scattered files.
Built: a connected workspace with team dashboards, daily reports in Notion, and leave requests with email approvals.
Result: every team member has a personal dashboard, and daily reporting moved off WhatsApp.
NotionSystems designHR workflows
Healthcare · 2022–2026Growing a Kolkata clinic: branch, launch, campaigns
Problem: a respected clinic wanted to reach more patients and add a new product line.
Built: opened a second branch, led an Emergency Kit launch (design, event, promotion), ran seasonal patient campaigns, and set up a Notion workspace and a patient-form dashboard.
Result: patient base up 20% in 6 months; seasonal campaigns grew monthly revenue 13% and 33% after breakeven.
LaunchCampaignsVendorsNotion
Side projectBenjamin, an AI operations assistant
Problem: daily reports and task follow-ups ate hours every week.
Built: a WhatsApp assistant that sends a daily task digest, reads reports and logs attendance, running partly on local AI models.
Result: a working prototype I use myself.
AI agentsNext.jsWhatsApp