Hoang Dang Tan Phat - AI / Senior Software Engineer

Professional Summary

Backend engineer with 8 years of experience (2017 – 2025) building AI-powered systems and scalable microservices. Specialized in developing production LLM applications using retrieval-augmented generation (RAG), LLM agents with LangGraph, and semantic search architectures. Skilled in designing backend systems that integrate Claude API, LangChain, vector databases (pgvector), and real-time services. Expert in Node.js/TypeScript, Python/FastAPI, Golang across distributed systems, microservices, and cloud infrastructure. Passionate about building conversational AI systems, AI-assisted workflows, and production-grade AI backends.

Technical Skills

AI & LLM

Backend Languages

Node.js/TypeScript (4 yr) | Golang (5 yr) | Python (2 yr) | NestJS | FastAPI

Databases & Search

AI Infrastructure

Whisper | MLX | Voyage AI Embeddings | Hugging Face | Real-time Streaming APIs

Architecture & Patterns

DevOps & Cloud

Frontend (Supporting)

Next.js | React | TypeScript | Tailwind CSS | Chrome Extensions

Observability & Monitoring

OpenTelemetry | Grafana | Prometheus | Loki | Elastic APM

Professional Experience

Caliber Venture (member of KyberNetwork)

Aug 2023 – Present

Optimex — Decentralized Bitcoin Trading Platform (Apr 2024 – Present)

Position: Backend Lead

Built Node.js/NestJS and Golang microservices with React frontend serving 2,000+ users, processing $4.9M+ trading volume across 2,247+ cross-chain transactions
Designed cross-chain architecture with Ethers.js for EVM integration, handling 100+ daily transactions across Bitcoin, EVM, and Solana networks
Built AI customer support agent using LangGraph + RAG to answer domain-specific questions about trading, lending, crypto, and DeFi protocols with semantic knowledge retrieval and Claude API integration
Set up production infrastructure on AWS/GCP with ArgoCD achieving 99.5% uptime
Reduced API response times by 60% through Redis caching and PostgreSQL query optimization

Tobi — AI-Powered Web3 Telegram Mini App Wallet (Aug 2023 – Apr 2024)

Position: Software Engineer

Built Node.js/NestJS and Golang microservices with React Telegram Mini App, serving 8,000+ users
Integrated Kafka event streaming processing 5,000+ events/day with guaranteed delivery
Implemented CDC (Debezium) for real-time data synchronization, reducing sync latency to sub-second

OnPoint Ecommerce

Aug 2020 – Jul 2023

Opollo — SWAT 2 (CRM & E-commerce Platform) (Aug 2020 – Jul 2023)

Position: Team Lead

Led team of 8 developers delivering full-stack platform handling 50,000+ orders/month
Built backend services with Elixir leveraging OTP for fault-tolerant, concurrent order processing
Integrated 3 payment gateways (VNPay, Momo, ZaloPay) processing 20,000+ transactions/month
Implemented Elasticsearch + Kibana for product search and operational analytics dashboards

Hermes — Multi-channel Chat SaaS Platform (Jan 2021 – Jul 2023)

Position: Key Developer

Built real-time chat platform handling 500,000+ messages/month across Lazada, Shopee, Facebook, and Zalo
Designed Golang microservices with gRPC achieving <100ms response time
Built NATS JetStream pipeline processing 20,000+ events/hour with at-least-once delivery
Developed Auto Reply system reducing customer response time by 60%

NUS Technology

Jun 2017 – Jul 2020

Whalesave — Financial Management Platform (Jan 2019 – Jul 2020)

Position: Team Lead

Led team of 5 developers through complete product lifecycle from concept to production launch
Built Express.js backend with MongoDB handling 10K+ requests/day
Integrated Stripe, Rapyd, and PayPal payment processors handling 1,000+ payments/month

Welligence — Data Analytics Platform for Oil & Gas (Jun 2017 – Jul 2020)

Position: Software Engineer

Led distributed team of 5 developers across Vietnam and United States time zones
Implemented Elasticsearch full-text search indexing 500K+ records with sub-second response
Built database sharding strategy supporting 5x data growth without performance degradation

AI & Featured Projects

Glow — AI-Powered E-commerce with LLM Customer Support

Full-stack skincare e-commerce platform with AI customer support powered by LLM agents and retrieval-augmented generation. Demonstrates production AI system combining conversational AI, semantic search, and real-time chat.

LangGraph multi-intent agent workflow with specialized handlers for order tracking, product search, skincare recommendation quiz, and FAQ retrieval
RAG pipeline using Voyage AI embeddings + pgvector for semantic knowledge base retrieval from product/support documentation
Python FastAPI microservice integrating LangChain + Claude Haiku with confidence scoring and human escalation fallback
Real-time streaming chat using Pusher + Next.js frontend with Prisma ORM and Better Auth
Stripe payment integration + admin dashboard for inventory & order management

InterAI — Chrome Extension + Python Backend for AI Meeting Assistant

Real-time meeting assistant integrating speech-to-text transcription, Chrome extension frontend, and LLM-powered assistance. Demonstrates real-time audio streaming, API orchestration, and browser extension development.

Chrome Manifest V3 extension capturing Google Meet audio via Web Audio API with low-latency streaming
Python FastAPI service with MLX Whisper for fast, on-device speech-to-text transcription
Keyword detection triggers Claude/OpenAI API calls for contextual meeting assistance
Live transcript rendering in extension side panel with real-time updates

Key Achievements & Impact

                AI System Architecture
                Designed LLM agent orchestration using LangGraph for multi-intent conversational workflows
Implemented RAG systems with semantic search for improved answer relevance and reduced hallucination
Built microservice separation of AI inference from core backend for independent scaling
Integrated confidence scoring and human escalation patterns for robust production AI

            

                Scale & Performance
                Processed $4.9M+ trading volume on decentralized blockchain platforms
Built systems handling 500K+ messages/month with <100ms latency
Achieved 99.5% uptime serving 2,000+ concurrent users

            

                Leadership & Team Growth
                Led engineering teams of 5–8 developers across distributed locations (Vietnam, US)
Delivered 6+ production systems from concept through launch and scaling
Mentored junior engineers, conducted code reviews, established best practices

            

                Performance Optimization
                Reduced API response times by 60% through caching and query optimization
Optimized background job processing, cutting execution time by 40%
Decreased incident response time by 70% with monitoring & alerting

            

Education

VNUHCM — University of Information Technology

Bachelor of Software Engineering | GPA: 7.9/10

2014 – 2019

Languages

Vietnamese: Native

English: Professional Working Proficiency