6+ years in performance marketing and growth. Now in M.S. Marketing Analytics at UMD, researching Hidden Markov Models, building Bayesian MMM in Python, and looking for roles where measurement directly informs decisions.
Six years in growth and performance marketing: paid media at scale, acquisition funnels, product launches across India and Bermuda. Somewhere in that work, I kept hitting the same wall: platform data couldn't be trusted. Every channel claimed credit. ROAS looked great on the dashboard and flat in reality. That question, how do you actually measure what works, is what pulled me toward analytics.
I'm also a builder. Started Caroma Fragrance from scratch in India, currently building an AI journaling product for mental health at UMD's Xfoundry accelerator. The 0-to-1 problem is the one I find most energising.
Now completing an M.S. in Marketing Analytics at UMD Smith, doing HMM retention research and building Bayesian MMM in Google Meridian. Trying to combine six years of campaign intuition with the statistical rigour to actually prove what moved the needle.
Built a full Bayesian MMM using Google Meridian (Python/PyMC) on 52 weeks of multi-channel campaign data. Specified hierarchical priors for channel ROAS, fitted posterior distributions via MCMC sampling, and interpreted response curves and saturation thresholds. Delivered budget reallocation scenarios with quantified uncertainty, answering not just "what worked" but "how confident are we."
Research assistant to Prof. Xinying Hao. Writing production SQL to extract, validate, and transform user-level behavioral data for a Hidden Markov Model quantifying how push notification timing and content influence app retention and churn transitions across hidden engagement states.
End-to-end automated pipeline: ingests campaign tracker data (Excel), performs multi-step validation and cleaning (pandas), produces structured performance views by channel and line of business, then connects to the OpenAI API to auto-generate AI-written executive summaries in branded PowerPoint decks. Saves 120 hours/year in manual reporting.
Founder at UMD's Xfoundry startup accelerator, building an AI-powered journaling product for mental health. Owns full go-to-market strategy: market entry analysis, competitive positioning, user persona development, and acquisition funnel design targeting underserved mental health demographics. Developed multi-year revenue projections and unit economics (CAC, LTV, retention curves), and designed lifecycle marketing strategy mapping user journey from first session through habit formation.
Looking for marketing analytics, growth analytics, or measurement roles where the work actually changes how decisions get made.