I build applied AI, and I point it at markets.
I’m a Management Engineering student at Waterloo. I like the part where a messy pile of text turns into something you can actually act on, and I put my own money behind the thesis, which is a good way to find out fast when I’m wrong.
I spent this summer building data and AI solutions for banking clients at Tiger Analytics in Santa Clara. This term I’m a research assistant in Waterloo’s SHARE Lab, working with Prof. Sharon Ferguson on whether the benchmarks we use to judge frontier AI measure what they say they do. I’m also an ML researcher at WAT.ai on InsightPulse, and co-founder of Click A Diet, an AI nutrition platform with ~200 paying customers. My interest sits at one intersection: applied AI, capital markets, and the discipline of putting real money into live ones.
Python (pandas, NumPy, scikit-learn) · SQL (window functions, CTEs, joins) · R · VBA
Regression · Model evaluation · Benchmark auditing · LLM integration · Evaluation harnesses
PostgreSQL · Supabase · Vercel · Git · Excel (advanced)
I work in the SHARE Lab at the University of Waterloo, led by Prof. Sharon Ferguson. We study the human-facing benchmarks used to judge frontier AI, the ones where people write the items or make the calls. The aim is to pin down what these benchmarks catch and what they miss, then build better ones.
I launched an AI service that emails subscribers a personalised 28-day diet plan every month, now reaching around 200 paying customers. Quality was the hard part. I built an evaluation harness that pairs each generated plan with a blind second-model recheck, lifting the pass rate from 5% to 89%. A verification layer sits in front of the customer and blocks any plan failing a calorie floor, a macro target or a condition-specific limit, which cut nutritional error by roughly 95%.
I turned transaction-level records into model-ready customer features, writing SQL and Python pipelines that leaned on window functions and CTEs over large banking datasets. That work fed a client's personalized-offer project: I assembled the customer dataset behind the logic deciding which offer each customer saw, and through which channel. Every deliverable shipped with its methodology, assumptions and findings documented, so a reviewer could trace the path from raw data to final output.
I built the team's baseline regression model in scikit-learn, mapping 8 macroeconomic indicators onto 6 market targets across 10 years of monthly data. The more useful result was a negative one: I found CBOE Volatility Index data leaking into the equity targets, which had inflated our test R² to +0.30. Removing it revealed the true value of −0.52, and that honest baseline is what the rest of the work now builds on.
My first time using data to drive operational decisions in a regulated industry. I wrote Excel VBA scripts that automated budgeting and cost-tracking during finance reviews, cutting about 20% off the monthly reporting cycle, then broke spend down across cost centres so the finance and operations team could see where costs were actually concentrated.
A cross-asset financial intelligence platform. Baseline regression mapping 8 macroeconomic indicators onto 6 market targets over a decade of monthly data, plus the leakage audit that kept the result honest.
A Python and Google Sheets investment engine that ingests live market data, applies portfolio constraints, and selects allocations across 10 asset classes.
A nutrition platform running two ways at once: AI-generated personalised diet plans and one-on-one expert dietician consults, under one roof.
Designed the survey and analysed responses from 355 teenagers for a peer-reviewed study on the health effects of digital gaming, with Dr. Vinay Goyal, Director of Neurology at Medanta, The Medicity.
A personal portfolio kept since June 2025. Real money, modest sums, tracked carefully. The exposure is ETF-driven and thematic: broad indexes, precious metals, US tech, and one targeted bet on defence. It is a small sum. The point is the process.
The base the rest sits on.
Carry for the tail.
The same trade, more beta.
Where the growth still is.
A policy cycle, not a whim.
At my size, any edge from individual stock picking gets eaten by execution friction and the urge to fiddle. ETFs let me express a thematic view with single-digit expense ratios, no idiosyncratic blow-up risk, and a process I can actually stick to.
Defence is in a structural reshoring cycle globally: multi-decade capex commitments, indigenization mandates across major economies, a long order book. It is policy-backed rather than cyclical, and the thesis plays out across European primes, US defence, or any clean expression of the trade.
Five sectors, ETF-driven. Not investment advice.
I’m looking for an Applied AI / ML Engineering internship for Winter 2027 at firms that take both AI and capital markets seriously. If that sounds like you, I’d love to talk.