Skip to content
View NasitSony's full-sized avatar

Block or report NasitSony

Block user

Prevent this user from interacting with your repositories and sending you notifications. Learn more about blocking users.

You must be logged in to block users.

Maximum 250 characters. Please don’t include any personal information such as legal names or email addresses. Markdown is supported. This note will only be visible to you.
Report abuse

Contact GitHub support about this user’s behavior. Learn more about reporting abuse.

Report abuse
nasitsony/README.md

Hi, I'm Nasit Sony 👋

Distributed Systems & AI Infrastructure Engineer

I build correctness-first distributed systems—from storage engines and consensus protocols to AI orchestration platforms.

My work focuses on one question:

How can distributed systems remain correct under crashes, retries, network delays, and adversarial failures?

Featured Projects

  • VeriProtocol — Scheduler-aware framework for consensus experimentation and protocol evaluation.
  • VeriStore — Crash-consistent storage engine with WAL, recovery, snapshots, and replication.
  • VeriFlow — Kubernetes-native AI workload orchestrator with GPU-aware scheduling.
  • LLM Serving Cache — Cache control plane for efficient LLM inference.
  • SmartSearch — Fault-tolerant semantic ingestion and retrieval platform.
  • AgentFlow — Workflow engine for reliable task execution and recovery.

Open Source

Contributor to:

  • TiKV (Rust)
  • etcd/raft
  • Temporal
  • Kubernetes SIG

Research Interests

  • Distributed Systems
  • Consensus Protocols
  • Storage Systems
  • AI Infrastructure
  • Reliability Engineering

Tech Stack

Languages: Rust, Go, Java, C++, Python

Infrastructure: Kubernetes, Kafka, PostgreSQL, Docker

Distributed Systems: Consensus, WAL, Replication, Scheduling, Recovery

📚 Google Scholar: https://scholar.google.com/citations?user=mBIQ1-0AAAAJ&hl=en

🔗 LinkedIn: https://www.linkedin.com/in/nasitsony

Pinned Loading

  1. VeriStore VeriStore Public

    Correctness-first C++ storage engine with WAL durability, crash recovery, Raft replication, and a minimal S3-style object store.

    C++

  2. veriflow-control-plane veriflow-control-plane Public

    Checkpoint-aware, fault-tolerant Kubernetes job orchestrator for AI training workloads

    Go

  3. SmartSearch SmartSearch Public

    Production-style semantic search and RAG backend built as a distributed system. Features async ingestion (Kafka), embedding pipelines, pgvector search, and strong reliability guarantees — including…

    Java

  4. async-bft-suite async-bft-suite Public

    Prototype framework implementing three asynchronous BFT agreement protocols (Cachin MVBA, VABA, pMVBA) with a unified simulation harness and comparable metrics.

    Python

  5. llm-serving-cache llm-serving-cache Public

    Distributed inference cache using VeriStore

    C++

  6. agentflow agentflow Public

    Control-plane system for reliable, stateful task orchestration with idempotency, retries, and failure-aware execution.

    Java