Ho Chi Minh City, Vietnam

Hi, I'm Leon Pham

Software Engineer

Software engineer building scalable backends and polished frontends. I work across the full stack — from APIs to UIs.

View writing Get in touch

AWS SA Certified2+ yrs expUseful for dev

Experience

Mar 2025 — Present

Software Engineer

Now

Oven

Project: Monitoring Platform · Tackman

Built a centralized monitoring pipeline for internal projects using Grafana, Alloy, Loki, and Prometheus, making it easier for teams to track metrics, logs, and slow APIs
Designed and implemented a video upload and processing workflow that converts videos to HLS format (m3u8) using AWS MediaConvert and automatically generates subtitles, enabling the system to reliably handle 30+ concurrent video uploads without data loss
Optimized the subtitle-to-audio conversion pipeline on AWS Lambda, reducing processing time by 50% and resolving timeout issues for long subtitle files (30–60 minutes)
Built an automated pipeline for generating video assets, including multi-language subtitles and corresponding audio tracks, reducing manual work and improving content production efficiency

redis

node

postgresql

prisma

docker

typescript

react

aws

May 2024 — Feb 2025

Fresher Backend Developer

ATOM SOLUTION

Project: Digistore BVBANK · Apache Fineract

Implemented service separation to reduce system complexity, improving startup time and optimizing server resource utilization
Implemented data synchronization between databases using Debezium and Kafka, enabling near real-time data consistency across services and reducing reliance on manual data syncing or batch jobs
Researched and integrated Keycloak for single sign-on (SSO), centralizing authentication across multiple services, reducing duplicated auth logic, and improving both security and user experience

redis

python

postgresql

docker

Open Source

Personal research and open source projects I build outside of work.

longtk26 / ecommerce-microservices

A distributed e-commerce system demonstrating the Choreography SAGA pattern for distributed transactions across Order, Inventory, and Payment services with automated failure rollbacks.

GitHub

Implemented Event-Driven Choreography SAGA to manage distributed transactions across Order, Inventory, and Payment services without two-phase locking
Engineered automated compensating transactions and rollback workflows (e.g., inventory restock on payment failure) ensuring eventual consistency
Eliminated overselling race conditions using JPA optimistic locking (@Version) and decoupled asynchronous messaging via RabbitMQ topic exchanges
Configured Spring Cloud Gateway with AWS Cognito JWT validation, token relay, and orchestrated deployment via Kubernetes Kustomize

JavaJava 21Spring Boot 3RabbitMQSAGA PatternPostgreSQLKubernetes

longtk26 / ai-chatbot

Research on methodologies to extend AI model capabilities — tool function calling, Model Context Protocol (MCP), and Retrieval-Augmented Generation (RAG).

GitHub

Implemented tool function-calling with OpenAI API to enable AI agents to invoke external services dynamically
Integrated Model Context Protocol (MCP) via Google Calendar MCP to connect LLMs with real-world data
Built a RAG pipeline with LangChain, pgvector, and OpenAI embeddings to ground responses in private data
Developed a multi-turn chatbot UI with Gradio to demonstrate tool-augmented LLM interactions

PythonOpenAILangChainRAGMCPpgvector