Vancouver, BC

A software engineer who sometimes ventures beyond software.

I'm a multidisciplinary engineering grad from UBC. I like to work on cloud infrastructure, backend services, embedded hardware, and applied ML. I like to have a bird's eye view of different parts of the software and/or product development lifecycle. Currently looking for software, infrastructure, or platform engineering roles. In my spare time, I am nerdy about electronics and a variety of other things.

☁️ cloud infrastructureπŸ€– roboticsπŸ”’ privacy & AI research🍳 culinary arts🎡 musicπŸ”§ hardware tinkering

A few things I've built

All projects β†’
Personal project Β· ongoing
ongoing2026

tfbot

Convention-aware Terraform change automation β€” a self-hosted bot that turns a plain-English infra request into a convention-perfect, ready-to-review pull request. The model proposes a validated intermediate representation; deterministic Go performs the edits.

Organized around one bet β€” the LLM proposes, deterministic Go performs. The model only ever emits a small, schema-validated IR; an hclwrite engine turns that into surgical AST edits, so existing comments, ordering, and hand-tweaks stay byte-for-byte intact.A zero-LLM, offline scanner reads a Terraform repo and drafts a recipe.yaml of its conventions β€” module-vs-raw, file placement, naming scheme, shared defaults β€” so one generic executor can treat new resource types as data instead of new code.
GoTerraform / HCLhclwrite ASTAnthropic Claude APIJSON Schema
Personal project Β· ongoing
ongoing2026

CogniShield - Vibe Check

A zero-LLM cognitive-dependency scorer that classifies programmer–AI chat transcripts into six behavioral archetypes using geometric signal-space gating and calibrated confidence bounds.

A dependency-scoring pipeline that runs end to end without a single LLM call at inference time β€” PII redaction, six-axis signal extraction, size-adaptive gist compression, and archetype-sphere gating.Adapts the midpoint-radius n-sphere formula from RAG retrieval research (US Patent 12,579,200 B1) into a selective-prediction gate: transcripts that land clearly inside one region get auto-labeled; ambiguous ones escalate to a human.
PythonNumPyscikit-learnStreamlitMatplotlib
IGEN 430 Capstone Project
2026

HAVEN / IGEN Shop Assistant

A smart tool vending system combining Raspberry Pi orchestration, Klipper motion control, FastAPI, React, MQTT, and RFID into a working physical prototype.

Full-stack electromechanical system from concept to demo: frontend, backend, machine orchestration, motion macros, RFID auth flows, and admin tooling.Designed bounded slot allocation and dispense/return state machines for six rotary tool cakes with encoder-based physical verification.
Raspberry PiFastAPIReactTypeScriptDocker

Where I've worked

Jan 2026 – May 2026

Machine Learning Privacy Research Assistant

University of British Columbia

Trained PyTorch shadow models and implemented membership inference attacks to evaluate ML privacy leakage. Improved low-FPR attack performance using statistical scoring and multivariate feature analysis across confidence, entropy, margin, and loss-derived features.

PyTorchPythonML Privacy

Sep 2025 – Apr 2026  Β·  Sep 2023 – Apr 2024

Teaching Assistant β€” Introduction to Engineering I & II

University of British Columbia Β· Vancouver

Two separate terms as TA for a first-year engineering studio course. Facilitated hands-on sessions and helped 150+ students survive the jump from high school to engineering. Graded deliverables and invigilated exams for 250+ students.

TeachingEngineering Education

Jan 2025 – Aug 2025

Cloud DevOps Engineer Co-op

Samsung Research Canada Β· Vancouver

Raised IaC coverage to 90%+ across 8+ GCP projects by building a Python Terraform discovery, drift detection, and import pipeline for 300+ unmanaged resources. Reduced Docker image size from 2 GB to 55 MB via multi-stage builds, cutting CI runtime by 65%. Engineered Kubernetes deployments for internal services using Helm and GitOps. Led incident response and authored RCAs for BigQuery, Grafana, and CI/CD failures.

TerraformGCPAWSKubernetesHelmDockerPython

May 2024 – Dec 2024

Software Developer & Technical Writer Co-op

Samsung Research Canada Β· Vancouver

Built a TypeScript VSCode extension with AI suggestions, broken-link detection, and custom Markdown linting. Developed a Python automation tool using Selenium, Pandas, and NLTK that cut partner request review time from 15 minutes to 5 seconds. Built frontend and backend for a test analytics platform using Vue.js, Flask, Docker, AWS DocumentDB, and Elasticsearch. Wrote knowledge base articles for the Samsung Knox B2B offerings.

TypeScriptPythonVue.jsFlask

Sep 2021 – Oct 2024

Software Developer

UBC Uncrewed Aircraft Systems Β· Vancouver

Joined in 2021 as a mechanical member with essentially no idea what I was doing. Spent most of that year asking questions, staring at the whole system, trying to figure out how a claw would retrieve an object mid-flight. Eventually moved to software. Built a TypeScript and Svelte ground control station frontend with live map view and drone telemetry. Added a QR scanner that sends competition mission parameters directly to the backend as JSON. Wrote Python scripts to stream video from a machine vision camera over TCP and UDP for real-time mid-flight object detection.

TypeScriptSveltePythonSolidworks

Sep 2021 – May 2026

B.ASc, Integrated Engineering

University of British Columbia Β· Vancouver

Computer and Electrical specialization. Courses in systems software, computer networking, cybersecurity, software design, AI/ML, and blockchain.

Notes from the work

All posts β†’