Van Dat

Software Engineer | Database

I'm Tran Van Dat. I have three years of practical experience specializing in backend development, with a strong focus on concurrency, performance optimization, and database tuning. I am adept at researching and quickly adapting to new technologies, as well as optimizing APIs to deliver outstanding performance. My passion lies in improving database efficiency, enhancing both storage utilization and query speed. I thrive in collaborative environments and actively share knowledge to achieve common goals. I am always ready to collaborate and share knowledge to achieve common goals.

"The simple thing is the best"

Python · Go · Java FastAPI · Django REST· Gin · Spring Boot PostgreSQL · MySQL · Oracle · SQL Server MongoDB · Redis Kafka · RabbitMQ Observability · CI/CD · Docker · Cloud/On‑prem WebSocket · WebRTC

Skills & Technologies

Programming

Python (async/io) Golang (goroutine, channel) Java (Spring)

Prioritize clear, testable code, and benchmark using ab/vegeta or k6.

Framework

FastAPIDjango GinSpring Boot

Practice: DI, middleware, rate-limiting, tracing.
Auth: JWT, session, OAuth2.

Database & MQ

PostgreSQLMySQL SQL ServerOracle MongoDBRedis KafkaRabbitMQ

Plan/Index: EXPLAIN/ANALYZE, partial/index-only, GIN/JSONB, BRIN, covering.
Schema: partitioning, migration zero-downtime, lock strategy.
Pool: pgbouncer, max_connections, queue wait.

Database & Performance Optimization

Strategic Partitioning & Indexing

Partition by time-range + hash to reduce hot index size.
Use partial indexes for common query states.

Connection Pooling & Backpressure

Limit concurrency, separate read/write pools, and set proper timeouts.
Apply backpressure at the service layer to prevent resource exhaustion.

Query Planning & Caching

Refactor N+1 queries, use index-only scans, and leverage materialized views for reporting.
Implement Redis caching with key patterns and dynamic TTLs.

Observability & SLOs

End-to-end tracing (traceID) from gateway → service → DB → MQ.
Set alerts based on SLOs (p95 latency, saturation, error budget).

Highlighted Projects

EmagicEyes

AI-integrated system for operation assistance and automated information extraction.

Backend: FastAPI (Python async), Redis cache, PostgreSQL (JSONB + GIN), Kafka event bus.
Designed asynchronous processing pipeline with backpressure and idempotent consumers.
Result: average processing time ↓ ~55%; race condition errors nearly eliminated.

Demo/Docs (coming soon) Source (coming soon)