I'm studying Econometrics & Mathematical Economics at LSE, with interests in causal inference and applied data science. I'm particularly interested in causal questions: why does X cause Y, who is affected, and what can we do with that knowledge? I want to bring econometric research designs together with machine learning to investigate these questions, understand how effects vary across people and settings, and turn evidence into useful decisions. (Below I've grouped a selection of my research, coursework, collaborative work and personal projects into three areas: causal inference, machine learning projects and some wacky side projects)
More broadly, I like data science because it lets us look at numbers, discover relationships, and use what we learn to inform decisions around the world. I'm interested in both the methods and the substantive problems they help us understand. My curiosity has taken me into multilingual language technology, environmental data and small tools built in response to everyday problems (and I am excited to explore new fields!). I'm open to collaborating on new research ideas in causal inference and applied data science. If these interests overlap with yours, or you have a question you'd like to explore together, feel free to email me.
| Project | Work |
|---|---|
| CROWN Acts & labour markets | Independent paper using ACS data and a triple-difference design to study hair-discrimination protections. |
| Econometrics Game 2026 | Team research on education incentives and fertility, using Python and Stata. Second place. |
| RD external validity / AIPW_RD | Frontend and R/Python package development for the researchers' extension to RD designs. Private research repository. |
| Pé-de-Meia: schooling & work | Research proposal on education incentives and adolescent labour supply in Brazil, with sample-construction and study-design code. |
| Single-stair housing | Building-permit data collection and exploratory analysis of housing rules and building height. |
| Family background & education | Independent UKHLS/BHPS exploration and multiwave household-data construction in Stata. |
| Ozone policy analysis | Environmental panels, atmospheric-inversion inputs and RD/DiD diagnostics. |
| Project | Work |
|---|---|
| Consumer complaint language | Ongoing NLP work on template reuse and reported relief in CFPB complaints. Developed from DS202W coursework, with feedback from Dr Ghita Berrada. |
| Kuwait air quality | Missing-data analysis, wind features, regression and spatial imputation for PM2.5 monitoring. Collaborative research led by Dr Ghita Berrada and Dr Huda Alrashidi. |
| Agentic code switching | English–Tamil and Malay–English generation rules, annotation resources and review tools. |
| SwitchF1 | An ASR evaluation tool I developed to measure language-switch preservation using precision, recall and F1, alongside word error rate. |
| C360 refresh pipeline | R code developed for future use to save runtime and simplify recurring data refreshes. Code only; no datasets. |
| Project | Work |
|---|---|
| Visa appointment bot | Personal automation prompted by competition travel. Public code demonstrates appointment-date filtering and scheduling with local snapshots; upstream project credited. |
| Time-dependent routing | Graph search through changing obstacles when waiting is not allowed. |
| Pokémon weather map | Group coursework combining data collection, geographic information and a creative weather-based mapping rule. |
Adaptive visual-field testing — an exploratory simulation framework and reading list on adaptive statistical methods and the testing experience.
Compact view
One square per contribution, with empty days removed.
