██████╗ ██╗    ██╗███████╗███╗   ██╗    ███████╗██╗███████╗██╗  ██╗███████╗██████╗
██╔═══██╗██║    ██║██╔════╝████╗  ██║    ██╔════╝██║██╔════╝██║  ██║██╔════╝██╔══██╗
██║   ██║██║ █╗ ██║█████╗  ██╔██╗ ██║    █████╗  ██║███████╗███████║█████╗  ██████╔╝
██║   ██║██║███╗██║██╔══╝  ██║╚██╗██║    ██╔══╝  ██║╚════██║██╔══██║██╔══╝  ██╔══██╗
╚██████╔╝╚███╔███╔╝███████╗██║ ╚████║    ██║     ██║███████║██║  ██║███████╗██║  ██║
 ╚═════╝  ╚══╝╚══╝ ╚══════╝╚═╝  ╚═══╝    ╚═╝     ╚═╝╚══════╝╚═╝  ╚═╝╚══════╝╚═╝  ╚═╝

visitor@owenfisher.dev:~$ whoami

Backend & Infrastructure Engineer

I build real-time data and ML systems that hold up in production, and I ship products fast when the clock is running.

now:AI Engineer intern @ Sendbird
before:Quant Developer intern @ RBC (250x faster factor backfills)
wins:1st nwHacks 2026 · 1st YC track, HackTech 2026 · 1st Cursor AI Hackathon
focus:real-time ML systems, stream processing, AI agents
stack:Python, FastAPI, Kafka, Postgres, Redis, PySpark, AWS
education:UBC Computer Science, Class of 2027
location:Vancouver, BC, Canada
View ProjectsGet in touch

cat experience.log

Experience

Jun. 2026 — Present Sendbird

San Mateo, California, United States

AI Engineer · Internship
  • Building agentic AI workflows with LangChain.
  • Backend engineering in Python.
Sep. 2025 — May 2026 Royal Bank of Canada
Quantitative Developer · Internship
  • Migrated legacy MATLAB factor-scoring pipeline to Python/PySpark on Databricks, cutting single-factor backfills from 3.25 hours to under 1 minute for a 250x+ speedup across 30+ years of financial time-series data.
  • Owned Python POC rebuild of legacy MATLAB batch preprocessing pipeline, preserving backward-compatible file outputs while adding caching and multithreading to make daily runs more reliable and 15-20x faster.
  • Built regression validation framework across golden datasets, 26 intermediate calculation columns, 150 factor-region combinations, and 500M+ score rows, giving the team confidence to replace the legacy MATLAB pipeline.
  • Built Parquet/Delta migration pipeline centralizing 3.1B+ rows of legacy research data from proprietary MATLAB/NAS formats into Databricks volumes and governed Delta tables for faster querying.
June 2025 — Present UBC Science Undergraduate Society
Frontend Developer
  • Built production Next.js/TypeScript pages and reusable components for student-facing workflows, room bookings, clubs, and events.
May 2025 — Present UBC Actuarial Science Club
Frontend Engineer
  • Led the club website build and shipped event/member surfaces in Next.js, React, and Tailwind CSS.
Jan. 2025 — Mar. 2025 Quantico Research
Software Engineer
  • Provisioned Terraform-managed AWS infrastructure for data ingestion, embedding storage, and investment-factor generation.
  • Built Python/NumPy risk-signal pipelines and automated walk-forward validation and stress testing with Pytest inside CI.

ls projects/

Projects

cat projects/flux.mdMar. 2026
service trace
$ docker compose up kafka postgres redis api workers
[adapters] coinbase connected; kraken connected
[kafka] market_trades topic online; partitions=6
[consumer:candles] writing OHLC windows -> PostgreSQL
[consumer:indicators] publishing RSI/MACD snapshots -> Redis
[api] GET /markets /candles /crypto /indicators healthy
[metrics] consumer_lag=0 api_p95=42ms dlq=0

Backend / data infrastructure

Flux

Real-time market data pipeline with exchange ingestion, Kafka fan-out, FastAPI read paths, Postgres persistence, Redis-backed indicators, and Prometheus/Grafana observability.

PythonFastAPIKafkaPostgreSQLRedisDockerPrometheusGrafana
  • Exchange-aware ingestion
  • Kafka consumer lag alerts
  • REST + WebSocket APIs
  • DLQ and smoke-test coverage

git log --contributions

GitHub Activity

... contributions

github.com/obro79 · loading latest calendar

> github
LIVE

ws connect coinbase

Live Markets

Connecting to Coinbase WebSocket...

cat contact.txt

Contact

Open to backend, infrastructure, real-time ML, and AI-agent roles. If you're a founder building in those spaces, I'd love to hear what you're working on.

mail: