Business Analyst & Consultant. MS Business Analytics @ UMass Amherst (GPA 3.9). 3 years delivering data-driven decisions for Fortune 500 clients at Cognizant. Now building at the intersection of business analysis and agentic AI.
See My Work
I'm a Business Analyst and Consultant with a background in data analytics, client engagement, and enterprise delivery. I believe the hardest part of any project isn't the analysis - it's figuring out what question you're actually trying to answer.
At Cognizant, I worked directly with client teams at Chubb, a Fortune 500 insurance company, combining data analytics with client-facing delivery - translating complex operational data into decisions stakeholders could act on. Analysis is only as valuable as the clarity it creates.
I recently completed my MS in Business Analytics at UMass Amherst's Isenberg School, deepening my skills in machine learning, financial modeling, and strategic analysis - building the toolkit that modern BA and consulting roles demand.
Actively targeting BA and Consulting roles where data analytics and client engagement intersect. I am also building toward the intersection of business analysis and AI - designing agentic pipelines that automate the manual workflows BAs spend most of their time on.
I define the right problem before chasing solutions - the most underrated consulting skill.
SQL, Python, Tableau - from raw data to insight without hand-waving.
I translate technical findings into boardroom-ready narratives for non-technical audiences.
3 years coordinating across engineering, ops, and business teams in enterprise environments.
A Business Analyst spends 2-3 hours every Monday pulling churn reports, scoring risk, and writing an executive summary before sending it to leadership. What if that entire workflow ran itself? I built an end-to-end agentic pipeline on Snowflake that queries 41,000+ customer records, scores each segment HIGH / MEDIUM / LOW risk, detects revenue anomalies, and uses Groq LLaMA 3.3 70B to auto-generate a board-ready executive summary - delivered via email every Monday at 8am with zero manual input.
A business user has a question - "which high-value auto claims breach our escalation policy?" - but the answer lives in two places: a policy PDF and a claims database. What if one system could read the document, query the data, and brief you like an analyst would? I built a multi-agent system where a Router Agent decides whether a question needs document retrieval, a SQL query, or both - then a RAG Agent searches real insurance filings (Chubb 10-K, NAIC regulations, ISO coverage forms) while a SQL Agent queries live claims data, and a Synthesis Agent writes the executive briefing.
Out of 30,000 cardholders, which ones will stop repaying next month? Banks write off billions annually to defaults they saw coming - but acted on too late. I built a classification model to flag high-risk customers before they default, giving credit teams a data-backed early warning system.
Across 150,000+ guest reviews, are Airbnb hosts in Rhode Island pricing their listings based on what guests actually experience - or just what they wish they were worth? I used VADER sentiment analysis to score review tone by listing, then mapped those scores against pricing tiers to surface where hosts are overcharging relative to guest satisfaction.
Block Inc. runs two very different bets simultaneously - a consumer payments app (Cash App) and a Bitcoin treasury strategy. Does combining fintech scale with crypto exposure make Block stronger or more fragile? I analysed ~10,000 Google Play reviews with VADER, benchmarked Block's revenue streams against competitors, and assessed the $255M regulatory penalty impact on long-term positioning.
Radiologists miss roughly 4% of brain tumors on first read - can a model trained on labeled MRI scans catch what a tired human eye misses? I built a Convolutional Neural Network classifier to distinguish tumor-positive from tumor-negative MRI images, focusing on minimising false negatives where the cost of error is highest.
Gold moves on fear, inflation, and dollar strength - but which macroeconomic signals actually predict its price, and how far ahead can you see? I tested whether a combination of treasury yields, USD index, oil prices, and market volatility could reliably forecast short-term gold price direction using regression and time-series models.
MS Business Analytics graduate, available now. If you have a complex problem and need structured thinking, let's talk.