cv

This is a description of the page. You can modify it in '_pages/cv.md'. You can also change or remove the top pdf download button.

Basics

Name Dang Truong
Label Software Engineer
Email truongminhdang26403[at]gmail[dot]com
Url https://minhdang26403.github.io
Summary An engineer aspiring to apply deep distributed systems knowledge to the challenges of large-scale LLM serving and inference.

Work

  • 2025.06 - Present
    Member of Technical Staff
    Nutanix
    C++, distributed systems, and storage systems.
    • Replication protocols
    • Resource Scheduling
    • Data Protection
  • 2024.05 - 2024.08
    Member of Technical Staff Intern
    Nutanix
    C++, distributed systems, gRPC, Python.
    • Implemented a distributed snapshot operation in the Network File System (NFS), increasing snapshot capacity from 1,200 to 7,200 files and reducing snapshot time by 55%.
    • Optimized the NFS's distributed lock service and gRPC communication overhead, reducing file I/O latency by 80%.
    • Collaborated with Control Plane and Core Data Path teams to identify performance bottlenecks, achieving a 65% increase in overall system throughput.
  • 2023.05 - 2023.12
    Software Engineer Intern
    Teradata
    C++, database optimization, SQL, Python.
    • Work on the Workload Management component of the Teradata SQL engine, enhancing resource allocation and optimizing query performance for diverse database workloads.
    • Develop a real-time performance monitoring tool in C++ to track resource usage of 10,000+ concurrent queries.
    • Achieve an 80% reduction in message passing overhead by implementing a concurrent LRU cache within the tool.
    • Optimize the tool's throughput to handle 100,000+ requests per second, facilitating efficient query optimization for large-scale systems.

Volunteer

  • 2022.08 - 2022.11

    Remote

    Software Engineer
    Vietnam Tech Society
    React, Node.js, MongoDB.
    • Collaborated with 5 engineers to design and develop Techsphere, a mentoring platform with 10,000+ users.
    • Increased users' satisfaction by 25% by improving UI and fixing major bugs in the feedback and scheduling system.
    • Reduced rendering time of filtering feature by 30% by optimizing 20+ React components and network traffic.

Education

  • 2021.08 - 2025.05

    Granville, OH

    Bachelor of Science
    Denison University
    Computer Science
    • Data Structures and Algorithms
    • Operating Systems
    • Distributed Systems
    • Database Systems
    • Computer Architecture
    • Computer Networks
    • Parallel Computing

Skills

Distributed Systems
Replication Protocols
Distributed Consensus
Resource Scheduling
Storage Systems
Distributed File Systems
Key-Value Stores
Cloud Storage
Database Systems
Programming Languages
Modern C++
STL
Boost
gRPC
Protocol Buffers

Languages

Vietnamese
Native speaker
English
Fluent

Interests

MLSys
Large-scale LLM Serving
Efficient LLM Inference
Distributed Training
Deep Learning Compilers and Runtimes
Programming Languages
Modern C++
Python
CPython Internals

Projects

  • 2023.03 - 2023.06
    Distributed Key-Value Store
    A distributed key-value store implemented in Go with Raft consensus protocol, database sharding, and lazy compaction.
    • Designed and implemented a fault-tolerant key-value storage system using the Raft consensus algorithm, maintaining data consistency and availability in the presence of failures and network partitions.
    • Leveraged Database Sharding techniques to enhance the performance and scalability of the key-value store by distributing data across multiple replicated state machines.
  • 2022.06 - 2022.08
    BusTub Database
    A distributed key-value store implemented in Go with Raft consensus protocol, database sharding, and lazy compaction.
    • Implemented several futures for the database system, including Buffer Pool Manager, Hash Index, Query Execution, and Concurrency Control.
    • Designed a thread-safe Buffer Pool Manager that moves physical pages back and forth from memory to disk, allowing the DBMS to store databases that are larger than the amount of available main memory.
    • Reduced the lock contention between concurrent threads in the buffer pool by developing a Parallel Buffer Pool Manager that holds multiple buffer pool instances, each with its own lock.
    • Decreased query response time by integrating a Hash Index (a disk-backed hash table) into the DBMS to achieve data lookup in constant time.
    • Optimized the time to grow or shrink the database by using an extendable hashing scheme to dynamically modify hash functions for the Hash Index.
    • Built query executors that support sequential scan, insert, update, delete, join, and aggregation operations using the iterator query processing model.
    • Enhanced the DBMS to handle multiple access requests by implementing a Lock Manager that ensures concurrent transactions access shared data items in the correct order and prevents possible deadlock.
  • 2022.05 - 2022.06
    Concurrent Web Server
    A concurrent web server implemented in C with thread pool and LRU cache.
    • Built a web proxy that reads and parses HTTP requests from a web browser, forwards them to web servers, reads servers' responses, and forward those responses to corresponding clients.
    • Scaled the web proxy to handle multiple concurrent requests by adding a thread pool, reducing the overhead of spawning a new thread/process for each new client.
    • Decreased the access time of recently-used web objects by integrating an LRU cache to the proxy.