Case Studies & Systems

Selected Engineering Work

A detailed exploration into production systems, high-concurrency domains, performance turnarounds, and frontend architecture across 9+ years of engineering.

Founder & Full-Stack Engineer · Live SaaS2024 – Present

Somm Scribe: Wine Journal & AI Recommendations

Primary Outcome:

Live in production with active users, Stripe subscription tiers, and personalized taste recommendations.

Conceived, architected, and shipped a live wine tasting journal and personalized recommendation engine. Built as a high-fidelity web app backed by a high-concurrency Go REST API, PostgreSQL database, and Google Cloud Run deployment.

The Challenge:

Building a consumer SaaS as a solo engineer requires balancing fluid mobile-responsive UX, secure tier-gated subscriptions, resilient data pipelines, and production infrastructure on a lean operational budget.

The Solution:

Engineered an end-to-end architecture: React/TypeScript frontend with zero UI bloat, Go REST microservice on Cloud Run, OpenAI integration for personalized wine recommendations based on tasting history, and hardened Stripe webhook lifecycle handlers.

100%Solo Ownership
AIRecommendations
100%Live SaaS Billing
< 150msAPI Response Time
ReactTypeScriptGo (Golang)PostgreSQLStripeOpenAICloud RunDocker
Somm Scribe: Wine Journal & AI Recommendations preview
Live SaaS Screenshot
HiBob · Senior Frontend EngineerSept 2025 – July 2026

Bob Finance: Safeguards & Refresh for Financial Models

Primary Outcome:

Eliminated customer data discrepancies and safeguarded multi-source general ledger synchronization.

Engineered the Model Data Refresh workflow in Bob Finance, allowing enterprise finance teams to safely preview and validate pending general ledger and dimensionality updates before applying changes to financial models.

The Challenge:

When upstream general ledger accounts or dimensions changed, refreshing financial models without a clear preview made it hard for finance users to understand what would change before applying updates.

The Solution:

Designed a multi-stage preview pipeline in the UI. Customers can inspect pending GL and account diffs, review added/modified/removed dimensions, and confirm changes before refreshing financial models.

100%Data Safeguards
E2EPlaywright Suite
0Silent Data Overwrites
HighTable Density
TypeScriptAngularPlaywrightFinTechData ModelingState Management
Bob Finance PipelineSafeguards
1
Upstream GL & Dimensionality

Detect pending external accounting updates

↓
2
Preview & Diff Impact

Review pending GL and account changes

↓
3
Safeguarded Refresh

Automated Playwright regression coverage

When I Work · Senior Full Stack DeveloperAug 2021 – Sept 2025

When I Work: Scalable Payroll & Technical Planning

Primary Outcome:

Shipped high-accuracy payroll processing while elevating team-wide planning, testing, and mentorship standards.

Led frontend and backend development for core payroll features using React and Go. Solved complex regulatory rules around multi-timezone shift normalization, Daylight Saving Time boundaries, overtime thresholds, and blended position pay rates.

The Challenge:

Payroll calculation is unforgiving: an hour shift during Daylight Saving Time transition or miscalculating a blended overtime rate across multiple job roles directly affects employee wages and employer legal compliance.

The Solution:

Spearheaded technical development plans defining strict database access patterns, API contracts, and defensive boundary tests. Implemented robust calculation rules in Go and intuitive calculation verification interfaces in React.

4 YearsEngineering Ownership
100%Calculation Accuracy
RTLTesting Standard
Cross-OrgTechnical Plans
ReactTypeScriptGo (Golang)PHPREST APIsReact Testing LibraryArchitecture
Rules Engine ScopeReact + Go
DST & Timezones

Edge-case normalization

Blended Rates

Multi-role overtime pay

Defensive Plans

API & DB contracts

React Testing Library

Team-wide UI testing

When I Work · Senior Full Stack DeveloperAug 2021 – Sept 2025

Debugging Delayed Notifications in a High-Volume Scheduling System

Primary Outcome:

During peak windows most notifications did not go through, or arrived 5–30 minutes late. After SQL pagination, they were generally processed within a minute of the configured threshold.

Investigated reports that missed clock-in and clock-out notifications were delayed or not sent. During peak shift windows, most notifications did not go through. The failure only appeared at production scale: a PHP cron job loaded too much shift and time-entry data, ran out of memory, and restarted.

The Challenge:

Customers reported that those notifications were missing or arriving late. During peak periods, most notifications did not go through; those that did could arrive 5–30 minutes after the configured threshold. The issue could not be reproduced in lower environments, which did not have production-scale data volume.

The Solution:

Production logs showed out-of-memory errors clustered around common shift start and end times. Rather than raising the process memory limit, the job was changed to paginate in ~10,000-record batches, with additional logging and a Grafana dashboard for failures.

5–30 minPeak Delay Before
<1 minTypical Processing After
~10kRecords per Batch
GrafanaJob Visibility
Production DebuggingScalabilityObservabilitySQLPHPGrafana
Notification JobPHP Cron
1
Peak Shift Windows

OOM errors clustered at common start/end times

↓
2
Unbounded Query

Full result set loaded into process memory

↓
3
~10k SQL Batches

Paginated processing plus Grafana visibility

Calabrio · Software Engineer – UIJuly 2017 – Aug 2021

Calabrio: Workforce Management Performance & Zero Hotfixes

Primary Outcome:

Transformed sluggish real-time monitoring into sub-180ms UI and reached 0 production UI hotfixes.

Diagnosed and eliminated severe rendering bottlenecks in the enterprise Real Time Adherence widget using React Profiler. Introduced Cypress into the CI pipeline, achieving complete automated regression coverage across critical workforce management workflows.

The Challenge:

The Real Time Adherence widget re-rendered massive DOM trees on every incoming agent status event, creating perceptible UI freezing of over 2 seconds per event and severely degrading supervisor efficiency.

The Solution:

Used React Profiler to pinpoint cascading re-renders, decoupled volatile websocket state from static table layouts, implemented memoization boundaries, and established comprehensive Cypress end-to-end integration tests in Jenkins CI.

< 180msRender Latency
0Production UI Hotfixes
4 YearsCareer Progression
CI/CDCypress Automation
ReactTypeScriptAngularJSJestCypressDockerPerformanceJenkins
< 180msRender Latency

Optimized from >2,000ms (-91%)

0Production UI Hotfixes

Across all of 2020 via Cypress CI/CD

Want to see the complete career chronology?

Explore 9+ years of progressive scope or download the full Resume PDF.