- **Reduced deployment time by 66%** by implementing a
[solution](https://github.com/apache/incubator-kie-kogito-operator/commit/175a6356c5474f2360ccb8ae835e0b9b2d653cf1) for deploying locally-compiled binaries onto
Kubernetes/OpenShift via command-line, **cutting average
deployment times from 45 minutes to 15 minutes**.
(**Kubernetes/GoLang** used for this and three below).
- **Eliminated 80% of manual configuration errors** by enabling
the Kubernetes operator to automatically fetch data from
deployed services and update configurations, **deprecating
legacy startup scripts and reducing overall startup time
by 40%**.
- **Improved application stability** by introducing startup
probes for legacy applications with longer boot times,
**resulting in a 50% reduction in startup-related failures
and downtime during production launches**.
- **Enhanced system reliability** by refactoring probes to
[assign default values](https://github.com/apache/incubator-kie-kogito-operator/commit/af4977af228ec8648be28779259d4552246b656f) dynamically based on deployed YAML
files and fixing reconciliation issues, **increasing probe accuracy by 30%** and preventing misconfigurations.
- **Increased CI pipeline efficiency** by rewriting the
**Jenkins (Groovy)** [nightly pipeline](https://github.com/apache/incubator-kie-kogito-pipelines/commit/4c83f1aecdea2c1ba2796b79839a90d4083dce88) to run in a GitHub PR
environment, allowing for automated testing of all
team-submitted PRs prior to merging, **reducing manual
intervention by 60%**.
- **Increased project reproducibility** by taking initiative to
write a [reusable GitHub parameters file](https://github.com/apache/incubator-kie-kogito-pipelines/commit/4c83f1aecdea2c1ba2796b79839a90d4083dce88#diff-7d2c018dafbccec859077d19bf1ade53ec9c7649f235528ce89f5632b109f7e6) for the pipeline,
**enabling 100% reusability** and ensuring consistent pipeline
setups across different environments.
<!--- RBC AML {{{ -->
{{% /resume/section %}}<!--- }}} -->
{{% resume/project name="AML Risk Analytics"
languages="Python, SQL, Tableau"
date="July 2025" show="true" %}}
{{% resume/section projects %}}<!--- {{{ -->
* Built an end-to-end AML simulation using **Python**, generating **9M+ records** across customers, transactions, and alerts to mimic real-world financial behavior and suspicious activity patterns.
* Wrote advanced **SQL (CTEs + joins)** to segment **high-risk customers**, calculate alert counts, and filter transactions over the past 90 days with aggregated exposure metrics.
* Engineered a custom **risk scoring model** in Python using transaction thresholds and alert volume to classify customers as Elevated or Critical risk.
* Designed **interactive Tableau dashboards** (Risk Heatmap, Alert Efficiency, Risk vs. Avg Amount) to visualize cross-country AML exposure and alert effectiveness.
* Normalized multi-currency data and integrated key AML metrics like **alert rate (%)**, **avg USD amount**, and **transaction volume** for actionable reporting.
- **Built KPI-ready metrics** (alert rate, avg USD exposure, transaction volume) to support AML performance reporting and enable cross-country risk comparisons.
{{% /resume/project %}}
<!--- RBC AML }}} -->
<!--- Rarity Surf {{{ -->
{{% resume/project name="Rarity Surf"
languages="Python, Django, JavaScript, React"
date="Oct 2021" show="true" %}}
date="Oct 2024" show="true" %}}
- **Developed a full-stack web application** to generate rarity
- [**Redesigned item generation system**](https://github.com/Kevin-Mok/gobcog/pull/5) for open source Discord game
built with **Python**, replacing 83k-line static JSON files with dynamic item generation, achieving a **99% reduction** in file size and reducing memory usage by **85%**.
- **Implemented modular item components** to enable over **152,000 unique combinations**, improving gameplay diversity and item quality.
* **Reduced reporting deployment time by 66%** by building a CLI-based solution to push compiled binaries directly into Kubernetes/Openshift clusters, accelerating turnaround for testing and data validation workflows.
* **Decreased manual configuration errors by 80%** by automating service discovery and dynamic config updates, aligning with AML’s goal of minimizing operational risk and improving data integrity.
* **Improved system reliability** during production launches by implementing startup probes for legacy services, reducing downtime and enhancing stability for automated monitoring/reporting pipelines.
* **Enhanced CI pipeline reproducibility and performance** by rewriting the Jenkins nightly pipeline to support automated PR-level testing with reusable parameters, improving report consistency across environments.
* **Collaborated cross-functionally** with developers and testers to maintain reliable infrastructure, echoing the AML role’s emphasis on stakeholder partnership for building robust reporting systems.
