Bridging corporate finance and enterprise-grade DataOps: I architect end-to-end ETL pipelines, train ML models, and ship LLM-powered apps that turn raw data into decisions executives actually act on.
I'm a Data Professional with a Master of Science in Data Analytics, combining a foundation in corporate finance with enterprise-grade DataOps and cloud ecosystems.
My expertise spans the entire data lifecycle — architecting robust ETL pipelines, optimising complex SQL architectures, deploying CI/CD workflows, and integrating LLM applications into production environments.
I thrive at the intersection of technical data engineering and strategic business decision-making, with hands-on experience across real estate analytics, financial planning, and AI development in USA, India, and Australia. I've presented findings to non-technical stakeholders and translated ambiguous business questions into data problems worth solving.
A full-stack financial analytics platform tracking 26 S&P 500 tickers. Ingests RSS news feeds daily, scores articles via the Gemini API, and surfaces insights across 9 analysis tabs: OHLCV overlay, auto-ARIMA 7-day forecast, Random Forest Watch/Hold/Avoid signal with P(up) probability, sentiment vs. price chart, sentiment-return regression, volatility, STL decomposition, rolling correlation, and a 5-year income statement. A per-ticker Gemini AI analyst briefing with streaming chat Q&A rounds out the platform — all backed by PostgreSQL, deployed on a cloud VPS with full CI/CD.
Which airline should a frequent flyer trust, and does it matter when they travel? Using 10.7 million flights across 8 airlines (2019–2024), this analysis gives a data-backed answer. Post-pandemic delays have not recovered to pre-COVID baselines. Delta is the most reliable carrier at 1.77 min average delay; JetBlue the worst at 13.30 min. Summer travel carries 1.84× the odds of a 15+ min delay. Built with rigorous statistical testing: Welch's t-test, two-way ANOVA, Tukey HSD, logistic regression, and Random Forest.
Has the U.S. energy grid actually gotten cleaner — and which states are driving it? No clean dataset existed, so one was built from scratch via the EIA API and web-scraped archive files across 50 states and 11 years. CO₂ emissions dropped 144.78 kg/MWh on average (p=1e-12), but the gains are uneven: OLS regression and ANOVA reveal that renewable energy share is the strongest predictor of emissions intensity, with sharp regional disparities that national averages obscure.
Can physical seed traits predict how much a soybean cultivar will yield before harvest? Across 40 cultivars, Random Forest and XGBoost models (both regression and classification, tuned via GridSearchCV) say yes. Thousand Seed Weight and grain count per plant are the dominant predictors — a finding that holds up in the crop science literature and gives breeders a measurable selection signal earlier in the growth cycle.
A live tabletop RPG campaign generates more state than a spreadsheet can handle — characters, inventory, quests, gear, and loot all have complex relationships. This MySQL database models that domain properly: 3NF schema across 5 tables, a many-to-many inventory bridge, and a deliberate indexing strategy (FK joins, text search, boolean flags, composite indexes for equipped gear). Gameplay queries cover character sheets, encumbrance, quest tracking, and loot extraction.
| Degree | Institution | GPA | Country |
|---|---|---|---|
| M.S., Data Analytics | Walsh College | 4.0 / 4.0 | USA |
| PG Program, Data Science & Business Analytics | University of Texas | 4.0 / 4.0 | USA |
| B.Bus., Finance & International Business | Griffith University | 4.28 / 7.0 | Australia |
Connection established. I'm currently open to Data Analyst, Data Scientist, and Data Engineering roles — full-time, remote-friendly, US-based. Reach out via any channel below; I usually reply within 24 hours.