# shresh

> developer · data tools. personal site of shresh at https://shresh.ca.

this is the full content dump of the site for LLMs and crawlers. for the navigational index see [llms.txt](https://shresh.ca/llms.txt).

## about
i'm a data scientist at RBC and a math + cs / bba double-degree student at the university of waterloo and wilfrid laurier.

on the side i build small, useful things — mostly data tools, scrapers, and web apps. most of them start as "i wish this existed" and end up on the internet a few weekends later.

currently into competitive valorant analytics, ebook discovery, and porting cool unix-y tools to windows.

## projects

### vlr-api
unofficial vlr.gg API for live + historical valorant stats — ratings, clutch %, first kills, team rankings, match data, all filterable by tournament/region. self-heals when vlr.gg breaks: catches the error, asks codex to find the new selectors, and re-points the scraper.
tags: python, scraping, codex-api, self-healing
- github: https://github.com/shreshth-s/vlr-api

### bookwormd
browse bestsellers and find free downloads. made anna’s archive easy to browse so you never have to pay for an ebook again.
tags: web, next, scraping
- site: https://bookwormd.com
- demo: https://x.com/shresh_data/status/2032903731218448815

### clicky-windows
windows port of @FarzaTV’s clicky. saw it, thought it was cool, built it for windows. cleanup + open-source pass in progress.
tags: windows, desktop, open-source
- github: https://github.com/shreshth-s/clicky-windows
- demo: https://x.com/shresh_data/status/2041955796439736753

### val-events-radars
pulls + updates valorant pro player data and renders comparable radar charts per event. fast, lightweight, deployed on streamlit.
tags: python, streamlit, data-viz
- live: https://val-events-radars.streamlit.app/

## experience

### Data Scientist @ RBC
*Dec 2025 — present · permanent part-time · hybrid*

### Data Scientist (Co-op) @ RBC
*Aug 2025 — Dec 2025 · hybrid*

### Data Analyst (Co-op) @ RBC
*Jan 2025 — Apr 2025 · hybrid*

- piloted a contextual multi-armed bandit + RL framework for offer control testing, optimizing personalized offer splits across client segments.
- automated tableau dashboards via snowflake drivers — killed the manual reporting cycle.
- analyzed multi-year distributions of avion points earned/burned to identify top cohorts and inform loyalty strategy.
- produced insights + custom reports for loyalty, cards, and finance teams — offer investment analysis, linked client segmentation.
- stack: python (pandas, spark), sql (teradata, trino, snowflake), jupyter.

## education
- **University of Waterloo** — Mathematics and Computer Science (2022 — 2027)
- **Wilfrid Laurier University · Lazaridis School of Business** — BBA, Business Administration and Management (2022 — 2027)

## stack
- languages: python, sql, typescript, javascript
- frameworks: pandas, spark, streamlit, next.js, astro, fastapi
- tools: snowflake, teradata, trino, tableau, cloudflare, git, docker

## contact
- email: shreshthsharma1904@gmail.com
- github: https://github.com/shreshth-s
- x / twitter: https://twitter.com/shresh_data
- instagram: https://instagram.com/5hresh
- discord: shresh.s

## music
live spotify data (now-playing + top tracks / artists / albums) at https://shresh.ca/#music — not inlined here because it changes constantly. fetch directly from /api/spotify/now-playing and /api/spotify/top?type=tracks|artists|albums&range=short|medium|long for current data.

## blog

### hello, world
*2026-04-28* · url: https://shresh.ca/blog/hello-world/
tags: meta

> kicking off the blog. why this site exists and what to expect here.

i finally pointed `shresh.ca` at something other than an MX record.

this is the part of the internet where i'll write about the things i build — usually a mix of:

- **data + scrapers** — apis i wish existed, dashboards i needed once, things i scraped because someone said it couldn't be done politely
- **valorant analytics** — radar charts, per-event stats, vlr.gg internals
- **dev tools** — small windows utilities, ports of cool unix-y stuff, codex experiments
- **the boring middle** — deploys, dns, why my cloudflare email still works

new posts drop in `src/content/posts/`. that's it, that's the cms.

if any of this is useful to you, [say hi on x](https://twitter.com/shresh_data).