Theme
Language 中文

OPEN FOR CONSULTING & COLLABORATION

Turning ideas into things that run.

Mathematical physics by training. I build systems across forecasting, trading, production ops, and community platforms — and I'm learning the data craft in public.

See projects Read the letter
  • Python · PyTorch
  • Next.js 16 · TS
  • Rust · Tokio
  • Solidity · viem
BASED
Thunder Bay / Toronto
NOW
TSFM forecasting on equities + TaskMarket Base-mainnet prep

— FEATURED WORK

Selected projects.

A founder-shipped proxy product, a TSFM forecasting arena on equities, a Rust rewrite of a Hyperliquid trading bot, and a production console for a modular construction factory. Different domains, same instinct — make it real, then make it measurable.

ALL PROJECTS
SHIPPING · REVENUE SINCE 2022

Tensor Proxies

FOUNDER · PRODUCT & ENGINEERING LEAD

Proxy infrastructure for data collection, price monitoring, ad verification, and automation workflows.

Residential IPs
1M+
Uptime
99.9%
Entry price
$8 / 25
  • JavaScript
  • Docker
  • nginx
  • systemd
Updated Apr 2026
IN DEVELOPMENT · INTERNAL USE

hypebot-rs

SOLE AUTHOR

Rust rewrite of a Hyperliquid quantitative trading bot — typed pipelines, serialized execution, runtime resilience.

  • Rust
  • Tokio
  • reqwest
  • tokio-tungstenite
Updated Apr 2026
ACTIVE BUILD · USDC ESCROW ON BASE

TaskMarket

FOUNDER · CORE DEVELOPER

Campus services marketplace for international students — post a task, get bids, settle in USDC on-chain via a non-custodial escrow contract on Coinbase Base.

  • React 19
  • Vite 7
  • TypeScript
  • Tailwind 4
Updated May 2026

— TOOLKIT

What I actually reach for.

The honest version of the stack — not what looks good on a CV, but what shows up across the projects above. Python for forecasting, TypeScript for product, Rust for the parts that have to be fast and correct, Solidity when the marketplace needs settlement rails.

DATA & ML forecasting, sports & finance, honest benchmarks
  • Python
  • pandas
  • NumPy
  • scikit-learn
  • PyTorch
  • Chronos-2
  • TimesFM 2.5
  • FlowState
  • Monte Carlo
  • Poisson
  • Club Elo
WEB & PRODUCT app builds + bilingual marketing surfaces
  • Next.js 16
  • React 19
  • TypeScript
  • Astro 6
  • Tailwind 4
  • Drizzle ORM
  • Prisma 6
  • NextAuth v5
  • shadcn
  • Base UI
SYSTEMS typed services + realtime + low-latency runtimes
  • Rust
  • Tokio
  • reqwest
  • tokio-tungstenite
  • Node.js 22
  • Express
  • Socket.IO 4
WEB3 non-custodial settlement on Base
  • Solidity 0.8
  • viem
  • ethers v6
  • Hardhat
  • Coinbase Base
  • USDC ERC-20
INFRA & OPS single-VPS shipping discipline
  • Docker
  • nginx
  • systemd
  • Cloudflare Workers
  • MySQL 5.7
  • dbt
  • GA4
  • GitHub Actions

— NOW

What's active this week.

A rolling "now" page in miniature — the threads that are getting commits, the projects where the next decision matters.

— WRITING

Recent notes.

Working notes from forecasting, Rust rewrites, and shop-floor product decisions. Written to be useful six months from now, not viral today.

ALL POSTS
· 12 min · Forecasting / Sports / Data Analysis

Does the American heat move the 2026 World Cup?

A weather × performance study on the first 81 matches. Humid heat visibly drains players and reshapes tactics — but the body tax is absorbed before it reaches the scoreboard. Seven interactive charts, all on real data.

Read
· ~5 min · Reasoning / Learning / Exercise

Train the body to train the mind

I half-remember reading that exercise — especially team sports like football — is good for the prefrontal cortex: you train judgment and attention while you move. The solid part (aerobic exercise improves executive function) has thick evidence; the 'ball games beat running' part is thinner, but the direction has research behind it. So learning isn't only a sit-down activity — move, ideally inside a game that forces you to think fast.

Read

— ABOUT

Physics-trained. AI-native. Shipping in public.

Four years of mathematical physics at the University of Toronto left me a specific reflex: Fermi-estimate first, write down the counterfactual before reaching for a model, and trust a simple correct answer over an elaborate one.

Right now I'm building production systems on the side — proxy infrastructure, time-series forecasting harnesses, on-chain escrow, a Rust trading runtime — and writing about what I learn along the way.

What I'm filling in: the data and AI engineering craft — Python, SQL, scikit-learn, PyTorch, and enough ML-ops to ship. The side projects are the receipts.

READ THE FULL STORY
BASED
Thunder Bay / Toronto
EDUCATION
U of Toronto · Math + Physics
OPEN TO
Consulting & collaboration