01 / HelloPortfolio / 2026
Hello, I’m Felix.
FelixHuang.
01AI Engineer
02Software Engineer
03Data Engineer
Building AI-powered products, from reliable data systems to responsive iOS apps.
↓A little more about me02 / Work↘
The work
so far.
From AI products and iOS apps to automated job alerts and data analysis.
01Dec 2024 — Present
Skillbloom · Vancouver, Canada
Full Stack Developer
Building an AI-powered language learning app, from backend microservices and cloud infrastructure to iOS architecture.
- Built Go services handling 100 QPS at p95 latency under 390 ms; automated CI/CD and container infrastructure supported 99.5% availability.
- Improved perceived iOS load times by 30% with SwiftUI, MVVM, SwiftData, and asynchronous loading.
Go · SwiftUI · SwiftData · LangChain · Terraform · Docker
02Feb 2025 — Jul 2025
Outlier · Remote, Canada
AI Trainer
Refined AI-generated frontend code and interfaces for visual quality and consistency.
- Rewrote frontend responses to improve layout, UI aesthetics, and visual consistency.
- Optimized Next.js and Tailwind CSS implementations and provided detailed feedback on code quality and semantic correctness.
HTML · CSS · JavaScript · Next.js · Tailwind CSS
03Oct 2021 — Jan 2022
Tencent · Remote
Data Analyst Intern
Analyzed Tencent Wesee user behavior, short-video performance, and advertising data.
- Queried large-scale click, watch-time, and advertising datasets to identify audience preferences and content trends.
- Cleaned and transformed data with Python and Pandas, then built interactive Tableau dashboards for management.
Python · Pandas · Spark · Kafka · Tableau
03 / Projects↘
Ideas into
practice.
A data platform and two products built through my work, with the problems, contributions, and results behind each.
01Job Sub
Personalized job alerts across company career pages.
- The problem
- Monitor fragmented ATS platforms and deliver relevant openings without sending duplicate alerts.
- My contribution
- Built configurable adapters for Greenhouse, Ashby, Workday, Lever, iCIMS, Oracle Cloud, and Workable; added LLM parsing and preference matching.
- The outcome
- Tracked 102 companies every five minutes, with DynamoDB idempotency controls preventing duplicate alerts under SQS redelivery.
02Real-Time Transaction Risk Platform
A replayable streaming data platform for transaction risk.
- The problem
- Keep transaction records and risk events consistent while handling duplicate, invalid, and late-arriving data.
- My contribution
- Built a PostgreSQL transactional outbox with Debezium and Kafka, plus a three-tier Delta Lake lakehouse using PySpark and MinIO.
- The outcome
- Made derived state rebuildable through replay; validated event-ID deduplication, schema contracts, and recovery with automated chaos tests.
Kafka / Redpanda · Debezium · PySpark · Delta Lake · Avro · Redis · XGBoost
05 / About↘
Curious by
default.
I’m Felix, a full-stack engineer based in Montreal and a Canadian permanent resident. I build AI-powered products across backend services, cloud infrastructure, and iOS, with a focus on responsiveness and reliable data.
I earned a Master of Electrical and Computer Engineering at the University of British Columbia (2022–2024), with an overall grade of 87%.
Before that, I completed a Bachelor of Computer Science at Dalhousie University (2018–2021), graduating with First-Class Honours and a GPA of 4.1/4.3.
Felix ↗