⛴️ Ferryman · Zero-Downtime Postgres Migrator
A Go CLI that orchestrates zero-downtime Postgres cutovers. It backfills in parallel ctid page ranges, streams live changes through logical replication, and only cuts over once replication lag clears its prechecks.
| Stack | Go · PostgreSQL · Logical Replication · Docker · GitHub Actions |
| Scale | Parallel backfill across ctid page ranges with live change streaming |
| Safety | Replication-lag gating, enforced DDL ordering, automated rollback to source state |
| Reliability | CI runs every migration against seeded Postgres instances |
| Repository | Lushenwar/Ferryman |
🛡️ Sentinel · AI-Boosted Incident Response
An incident-response engine that cross-references git history against error logs to find the commit that broke production, then writes the diagnosis and postmortem for you.
| Stack | FastAPI · Next.js · TypeScript · PostgreSQL · RAG |
| Performance | Faulty commit surfaced as top candidate in ~13s; root-cause diagnosis in ~14s |
| Coverage | DB failures, memory leaks, null pointers, and other bug classes |
| Automation | RAG over runbooks; postmortems auto-posted to Slack when error rates cross thresholds |
| Repository | Lushenwar/Sentinel |
I'm a Math (Data Science) student at the University of Waterloo focused on infrastructure and site reliability engineering. I like the systems that keep everything else running: database migrations that can't lose data, incident tooling that finds the bad commit before a human does, and real-time code with bounded, predictable behavior.
My background spans embedded C++ flight software, a production clinical portal serving 400+ staff, and multi-agent AI pipelines. The common thread is correctness under load and knowing exactly what happens when something fails.
Open to: infrastructure, SRE, and backend engineering internships.
Languages
Frameworks
Data, Cloud & Infrastructure
| Domain | Proficiency | Details |
|---|---|---|
| Infrastructure / SRE | ●●●●○ | Zero-downtime migrations, logical replication, incident response, CI gating |
| Backend Systems | ●●●●○ | FastAPI, Express, PostgreSQL schemas, role-based access control |
| Systems Programming | ●●●●○ | Real-time C++, no dynamic allocation in control loops, build optimization |
| Applied AI | ●●●○○ | RAG pipelines, multi-agent LangGraph orchestration, K-Means clustering |
Software Engineer Intern · Mon Sheong Foundation
May 2026 – Aug 2026
Digitized fully paper-based care workflows into an end-to-end clinical portal.
- Built the Next.js / Node.js portal, cutting patient record retrieval time by 40% across daily nursing operations.
- Engineered role-based access control for Express APIs and PostgreSQL schemas, scoping every query by role for 400+ staff across 30 roles.
- Built real-time React dashboards with automated threshold-triggered alerts for clinical metrics.
Next.js Node.js Express PostgreSQL RBAC React
Embedded Flight Software Developer · Waterloo Aerial Robotics Group (WARG)
Sep 2025 – Jan 2026
- Removed dynamic memory allocation from the real-time control loop of C++ flight modules for deterministic, bounded execution.
- Restructured header dependency trees and build scripts, reducing firmware compile time by 25%.
- Optimized IMU sensor fusion filters to eliminate jitter, reaching 95% signal consistency.
C++ Embedded Real-Time Systems Build Systems Sensor Fusion
Research Lead · UpNxxt (via Riipen Labs)
May 2026 – Jun 2026
- Led UX research for the AI Sandbox, analyzing competitor engagement loops to cut onboarding friction.
- Designed child-first mechanics adopted into the roadmap, contributing to a 15% DAU lift over six months.
UX Research Product Strategy Retention
| Recognition | Details |
|---|---|
| 🥇 1st Place, GenAI Genesis 2026 | Eco-Pulse: urban heat planner over 178K land parcels |
| 🏆 Winner, DeerHacks V | LOCATR: 6-agent LangGraph venue planner, <1s cached |
| 🏆 Best Use of Snowflake | CxC 2026 |
| 🎖️ Finalist, 4 Tracks | YHacks Spring 2026 |
| 🥉 3rd Place, Beginner Stream | Hack404 |
learning: [ distributed systems, database internals, observability ]
building: [ zero-downtime data tooling, incident response automation ]
exploring: [ Kubernetes, SLOs and error budgets, chaos engineering ]
open_to: [ infrastructure / SRE / backend internships ]
Build systems that fail safely.