AI & Automation Developer

Nasdaq Verafin

The Mission

I am currently designing the architecture for an internal "Mixture-of-Experts" (MoE) service. The goal is to orchestrate specialized LLM agents that can autonomously review code for quality assurance and perform risk assessments before deployment.

This system moves beyond simple chatbots by integrating directly into the CI/CD pipeline, providing actionable feedback to developers in real-time.

src/agents/FeedbackAgent.java

public class FeedbackAgent extends MoEService {
  private final String MODEL =  "WMK-4-Sher";

  public Feedback generateReview(CodeContext ctx) {
    // Orchestrating specialized agents
    var riskScore = RiskAgent.evaluate(ctx);
    if (riskScore > 0.8) {
      return new Feedback("High Risk Detected");
    }
  }
}

Visuals

System Architecture

Details

DateSep 2025 - Present
LocationSt. John's, NL

Tech Stack

PythonLangChainAWS BedrockDockerGrafanaJava

Key Outcomes

Designed MoE Architecture
Orchestrated LLM Agents
Implemented SPACE Metrics
Cloud-native Analytics