Backend & Distributed Systems Engineer (Go/Java)

Nasit SSony

CA, USA

***********@*****.*** +1-702-***-**** linkedin.com/in/nasitsony github.com/NasitSony Summary

Backend / Infrastructure Engineer focused on building fault-tolerant distributed systems. Experience building Kubernetes-based job orchestration, async data pipelines (Kafka), and crash-consistent storage systems (WAL, Raft). Focused on correctness, failure recovery, and production-grade system behavior under real-world failure scenarios

TECHNICAL SKILLS

Languages: Go, Java, C++, Python

Infrastructure: Kubernetes, Kafka, PostgreSQL, Docker Backend: Spring Boot, REST APIs, Distributed Systems Concepts: Fault tolerance, WAL, idempotency, consensus (Raft, BFT) Selected Projects (Distributed Systems & Infrastructure) Veriflow - Kubernetes Job Orchestration Control Plane. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

(Go, PostgreSQL, Kubernetes, Docker) Feb 2025 - current, CA

- Built a Kubernetes-based job orchestration control plane with API-driven job submission and persistent life- cycle tracking in PostgreSQL

- Implemented concurrency-safe scheduling using PostgreSQL FOR UPDATE SKIP LOCKED, enabling multi- ple schedulers without duplicate execution

- Dispatches workloads as Kubernetes Jobs and reconciles runtime state into persistent storage

- Designed retry and backoff system with idempotent job submission and durable run-attempt tracking

- Developed reconciliation loop to sync Kubernetes Job state (RUNNING/SUCCEEDED/FAILED) into database after failures or restarts

SmartSearch - Fault-Tolerant AI Ingestion & Retrieval System. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

(Spring Boot, Kafka, PostgreSQL (pgvector), Spring AI, Docker)

Oct 2025 - current, CA

- Designed asynchronous ingestion pipeline using Kafka with idempotent processing and retry handling

- Implemented job lifecycle state machine (PENDING -> PROCESSING -> READY/FAILED) with deterministic recovery

- Handled failure scenarios including worker crashes, message replay, and database outages

- Built DLQ and retry logic to isolate failures and prevent infinite processing loops

- Ensured no duplicate embeddings under at-least-once processing using idempotency guarantees VeriStore - A correctness-first replicated storage system in C++. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . C++ Sep 2025 - current, CA

- Built a thread-safe key-value storage engine with append-only Write-Ahead Log (WAL) and fsync-based durability guarantees

- Implemented deterministic crash recovery via ordered WAL replay with corruption detection and safe trun- cation of partial log records

- Implemented Raft-based consensus replication with leader election, majority commit, log replication, and follower catch-up for fault-tolerant replicated state machines

- Designed snapshotting and log compaction to reduce restart latency and bound WAL growth

- Added group commit batching to reduce flush frequency and improve write throughput under concurrent workloads

- Built a minimal S3-style object storage layer on top of the KV engine supporting bucket creation, chunked object storage, metadata-based commit semantics, prefix-based object listing, and garbage collection of un- reachable object chunks

- Benchmarked durability strategies under write-heavy workloads (10k sequential PUTs), achieving 2.6x through- put improvement (104k ops/sec vs 39k ops/sec immediate flush) API Security Gateway for Microservices. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Spring Cloud Gateway, Keycloak, Docker, JWT Sep 2025 - current, CA

- Built a cloud-native API gateway with OAuth2/JWT authentication and role-based authorization for client and service-to-service communication

- Implemented rate limiting, request validation, and circuit breaker/retry policies to improve resilience and security under load

Async BFT Suite - Asynchronous Byzantine Fault-Tolerant Consensus. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Python, gRPC Jan 2025 - May 2025, CA

- Designed and evaluated prioritized multi-valued Byzantine agreement (pMVBA) under adversarial condi- tions

- Benchmarked against existing protocols, achieving up to 180% throughput improvement and 61% latency reduction

EXPERIENCE

University of Texas, Dallas. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Research Associate Sep 2025 - Current, Dallas, Texas

- Developing decentralized Byzantine-robust federated learning framework with formal convergence and fault- tolerance guarantees for decentralized AI systems. University of California, Merced. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Graduate Research and Teaching Assistant Aug 2018 - Aug 2025, Merced, CA

- Designed and evaluated distributed protocols improving throughput by 180% and reducing latency by 61% in distributed systems under failure scenarios.

- Built service prototypes simulating real-world network delays, batching, quorum coordination, and recovery strategies

- Published first-author research on Byzantine fault-tolerant consensus and atomic broadcast. Samsung Bangladesh R & D Center. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Senior Software Engineer Nov 2013 - April 2015, Dhaka, BD

- Optimized performance-critical C/C++ components, achieving 30% performance gain and 25% crash reduc- tion through profiling and concurrency debugging.

AtomAP Ltd.. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Senior Software Engineer Feb 2017 - Dec 2017, Dhaka, BD

- Built Java REST services for an international money transfer platform supporting 17 currencies; designed transaction-safe systems with ORM-backed relational databases for financial compliance workflows.

- Optimized SQL queries and service orchestration for high-volume, low-latency payment processing. Additional Industry Experience . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Senior Software Engineer December 2011 - June 2018, Dhaka, BD

- Developed iOS media-processing applications and backend-integrated web systems, improving performance, memory efficiency, and API scalability across consumer and enterprise products. PUBLICATIONS & RESEARCH IMPACT

- First-author publications on scalable Byzantine fault-tolerant consensus and atomic broadcast protocols with focus on cryptographic efficiency and distributed correctness. EDUCATION

University of California, Merced Merced, CA

MS in Electrical Engineering & Computer Science, GPA: 4.00/4.00

Contact this candidate