Hritik
Kumar

Senior Performance Marketing Lead

Ad accounts leak money. I find where, then build the tool that stops it.

On Google and Meta since 2021. Now $22K+ a day across a 250K+ SKU catalogue. I trained as an engineer, so when reporting disagrees with reality I don't file a ticket, I build the fix: an attribution layer that cut a 28% discrepancy to under 5%, and a survival model that predicts what gets reordered and when.

See the case studies

The gap I closed in reporting

Ad platforms and store data disagreed on revenue. Drag or toggle to apply the regression-based reconciliation layer I built.

Ad platforms reportStore confirmsUnmatched revenue

Reconciliation applied

~28%cross-platform reporting discrepancy

Revenue both sides agree onRevenue only the platforms reported

$22K+ad spend managed per day, about $680-700K a month

250K+SKUs modelled and managed

400%ROAS anchor, with an 800%+ peak over one 30-day window

3people led, across two brands

Case studies

Open a case for the problem, the build and the result. The waste-control case has a working version of the exclusion rule you can try.

Skills

Platform tenure, drawn to scale from 2021 to today.

Google and Meta start at 2021; the exact month isn't marked. Microsoft Ads is almost a year and now carries most of my managed spend.

Analytics

Power BI (DAX, drill-through)Power QueryExcelSQL

Modelling

PythonpandasXGBoostSHAPSurvival analysisR (clustering)

Systems

HubSpot CRM pipelinesShopifyFeed and tracking diagnostics

Way of working

AI-assisted building, with the logic, validation and data-leakage checks kept in my hands.

How I build

AI gets me to a working first version fast. The checking is what makes it worth trusting. Step through it, or use the arrow keys.

Phase 1 means a working, checked first version, not a finished product. I say so up front, then improve it against live results. Built this way: the reorder model, the reconciliation layer and the SKU-matching pipeline.

Path

From electrical engineering to running seven-figure ad accounts.

January 2026 to presentSenior Performance Marketing Lead, Mark Team AITwo B2B dropshipping brands selling to hospitals, clinics and universities.
January 2025 to January 2026Marketing Project Manager, FuaarkLed a team of 8 across paid media, SEO, creative and photography.
University of LeedsMSc Consumer Analytics and Marketing StrategyResearch assistant on candidate and marketing-professional studies, using clustering in R.
October 2022 to May 2023Marketing Manager, CauseIncBuilt the HubSpot pipeline from lead to MQL to SQL.
June 2021 to August 2022Digital Marketing Manager, Indiano PelletteriaAlongside a digital marketing internship at Kraftshala.
Where it startedB.Tech, Electrical and Electronics EngineeringThe origin of the debugging habit.

Let's talk

Open to senior paid media and growth roles, remote or hybrid.

hritikkumar27@gmail.com