Hi, I'm B. Sivasagar.

The problem. A household with a home loan, a car loan and a personal loan gets a bonus. Which loan should it go to? Bank calculators answer for one loan at a time and ignore prepayment charges, so people guess.

Key decisions.

1. Scoped v1 to 3–4 loans and three prepayment styles: extra each month, extra EMIs per quarter or year, or a fixed yearly lump sum.
2. Modelled Indian fine print: prepayment charges as a percentage or a fixed fee, with GST.
3. Shipped two formats, a web planner and a Google Sheets / Excel version, because users split between "just show me" and "show me the formula."
4. Launched one product first and held the follow-on (where spare money should go if not prepayment) for a second release.

What users told me. Early testers found it useful but hesitated to pay, since free calculators and AI tools exist. Rather than cut the price, I'm adding features that justify ₹199.

Brand. Dropped a crowded first name after checking every channel, and chose MapMyRupee so the handle is identical on Instagram, LinkedIn and Facebook.

The problem. Friend-group and league auctions run on a spreadsheet, a phone timer and someone shouting bids. Purses get miscounted, and teams overspend early and can't fill their squads.

Key decisions.

1. Rules are agreed up front and lock once bidding begins: participants, purse, bid timer and squad rules such as minimum bowlers and all-rounders.
2. Tiered bid increments that mirror real auctions (for example +₹5 L up to ₹1 Cr, then +₹10 L), configurable by the organiser.
3. Purse protection: a team can't spend so much that it can't complete its minimum squad, and the app explains why the bid button is disabled.
4. Bring your own players: a forgiving Excel/CSV import that understands "2 Cr", "50 L" and role short forms, flags bad rows instead of failing, and checks there are enough players for every team before you start.
5. Scoped v1 as one human team against simulated rivals, to test the feel of a live auction before building full multiplayer.

The problem. Each supplier batch arrives as images with prices and fabric details buried in text overlays and chat messages. Uploading by hand was slow, and inconsistent tags quietly broke the store's collections.

Key decisions.

1. Human in the loop, not full automation: for each product I choose whether the description comes from a category template, text read from the image, the WhatsApp message or my own notes.
2. Rules as requirements: a fixed SKU convention (vendor, fabric, work type, date), pricing rules, and a closed list of product types and tags that matches the store's collection rules.
3. Two paths by batch size: small batches go straight into Shopify through its API; batches of 50–100 go through a CSV import.
4. Consistency at the source: automatic photo normalization so products from different suppliers look uniform.

Along the way. Traced a recurring Shopify import error to a missing product category when certain attributes are filled, and made the tool warn before it happens.