Nandit Vaish

WORK SAMPLES / OCTOBER 2026

Useful, before elaborate.

Two deliberately small demonstrations. Inputs are fictional; these are examples of the method, not client delivery or outcome claims.

Engineering budget estimator

A workbook with Summary, WBS, Costs and Rates. Hours multiply the selected year and rate version. Expenses, fixed subconsultant fees, contingency and optional tax roll into the estimate.

Missing or ambiguous rate keys produce INCOMPLETE instead of an understated total. Append new rate versions rather than overwriting published rates; freeze accepted estimates as separate copies.

Illustrative USD planning fees, not salaries or construction costs. The static baseline is not protected against manual edits. If importing into Sheets, check the dropdowns and reconciliation before use.

Download the sample workbook

Contact cleanup with a paper trail

Fourteen invented rows become thirteen after one fully matching normalized record collapses. Ambiguous shared identifiers stay separate for review. Missing values stay visible; no enrichment or country-code guessing fills the gaps.

The sample includes before/after CSVs, a field change log, duplicate decisions, review flags, source-row lineage and the Python script. It demonstrates conservative processing, not mailbox validity, identity resolution or deliverability.

Import CSV contact columns as text to preserve phone numbers and IDs. The script requires Python and pandas. Test output before using it with your own records.

Download the cleanup example

Want a working sales example?

The separate Lantern CRM uses fictional solar enquiries. Assign an owner, set follow-ups and inspect the history. Browser profile switching illustrates responsibilities; it is not authentication.

Open the fictional CRM demo