Engineering Manager & Founding Team @ PhoenixDX Vietnam

Role Overview

As a founding team member at PhoenixDX Vietnam, I am building an AI-first engineering organization from the ground up — leading the entire Vietnam team and establishing the AI engineering culture. The mandate: re-imagine software delivery around autonomous AI agents, where the team operates as orchestrators of intelligent systems rather than line-by-line coders.

A Team of AI Experts

This is not a typical engineering team. Every member is a senior AI expert with 10–20 years of experience, recruited from top-tier organizations — NAB, SAP, MoMo, Binance, Pyco, and VinAI. I lead a hybrid org of AI-skilled engineers (who wield AI as a force multiplier) and dedicated AI engineers (who build the agentic systems themselves).

Multi-Agent Systems (MAS)

• ▹Architecting Multi-Agent Systems (MAS) for autonomous software delivery — moving beyond Single-Agent Systems (SAS) to coordinated agent fleets
• ▹Agent Orchestration: planning, task decomposition, and pipelining specialized agents (research, code, review, verify)
• ▹Agent Conflict Resolution: arbitration strategies when agents produce divergent or competing outputs — voting, adversarial verification, and consensus protocols
• ▹Agentic AI Pipelines: end-to-end automation chains that take work from intent to merged, verified code

AI-Driven Digital Transformation

• ▹Legacy migration at scale — 100+ legacy repositories re-architected into clean microservices, fully driven by AI agents
• ▹Frontend modernization: Angular 7 → latest React, automated by AI codemods and agent review loops
• ▹Elasticsearch-based rule management for dynamic, searchable business logic
• ▹RAG knowledge base — turning scattered tribal knowledge into a queryable, always-current source of truth

# PhoenixDX — AI-First Engineering
team:
  composition: "senior AI experts (10-20 yrs)"
  origins: [NAB, SAP, MoMo, Binance, Pyco, VinAI]
  model: "AI engineers + AI-skilled engineers"

multi_agent_system:
  paradigm: "MAS over SAS"
  capabilities:
    - agent_orchestration
    - agent_conflict_resolution
    - agentic_pipelines
  verification: "adversarial + consensus voting"

digital_transformation:
  legacy_migration: "100+ repos -> microservices (AI-driven)"
  frontend: "Angular 7 -> latest React"
  knowledge: [Elasticsearch rule mgmt, RAG knowledge base]

Building an AI Culture

Beyond the technology, the harder problem is culture. I am building a team that treats AI as a first-class engineering primitive — measuring throughput in shipped outcomes, not hand-written lines; trusting agents under rigorous verification; and continuously raising the bar on what a small team of experts can deliver. The goal is an org where humans set direction and verify, and agents do the heavy lifting.