name : Mayur Bhusare
role : Backend & Distributed Systems Engineer
location : Pune, India
status : GoQueue live - 2.9k req/sec, zero job loss
focus : Systems that survive real-world failure
DSA : LeetCode Knight | Contest Rating 1960+ | Top 3.05% GloballyI don't just write backend code - I engineer reliability.
My obsession is understanding why systems fail and building primitives that make them recover.
- Deep into queues, retries, DLQs, idempotency
- Building production infra in Go
- Heavy focus on observability and failure modes
- Always learning beyond coursework
Languages: Go (Golang), JavaScript, Java, Python
Backend & APIs: Go, Node.js, Express.js, REST APIs, Microservices-oriented design
Distributed Systems: Redis, RabbitMQ, Concurrent Worker Pools, Retry Orchestration, DLQ, Visibility Timeout, Async Processing
Observability: Prometheus, Grafana, Structured Logging
Reliability: Fault Tolerance, Idempotency, Rate Limiting, Graceful Shutdown, Backpressure, Load Shedding
Databases: PostgreSQL, MongoDB, MySQL
Cloud & Infra: Docker, AWS (EC2, S3, RDS, Lambda), Linux
Testing & Tools: k6 (Load Testing), Git, Postman
Fundamentals: System Design, OS, DBMS, Computer Networks, OOP
Go Redis PostgreSQL Docker Prometheus Grafana k6
Architected a distributed background job engine in Go (inspired by Sidekiq/BullMQ) using Redis for high-throughput queueing and PostgreSQL for durable job state; supports fan-out, priority queues, and deferred scheduling. Built concurrent worker pools with goroutines: exponential backoff retry, DLQ routing, visibility timeouts, graceful shutdown, stuck-job recovery via background reaper, and backpressure via queue-depth monitoring. Instrumented Prometheus/Grafana dashboards (queue depth, throughput, retry rates, worker utilization, p50/p95/p99 latency); k6 load-tested to 10k concurrent submissions at ~2.9k req/sec with zero crashes or job loss.
Node.js RabbitMQ Docker MySQL k6
Designed an asynchronous notification pipeline for Email/SMS/Push channels using RabbitMQ fanout exchanges with per-channel retry handling, DLQ isolation, and idempotent consumers. Sustained p99 delivery latency under 450ms under load testing with zero message loss across all retry and failure scenarios.
Node.js Express PostgreSQL Redis
Decoupled background worker ingesting 100+ job listings daily across 40+ CS roles without blocking the request path. Redis caching cut DB query latency from ~500ms to ~8ms (98% reduction). Enforced per-IP rate limiting at 60 req/min to prevent abuse under high-concurrency campus-scale traffic.
Node.js Redis BLPOP Semantic Versioning
Published a Redis-backed queue library with retry orchestration, DLQ routing, idempotency guards, and graceful shutdown, available on npm with 500+ downloads. Managed semantic versioning and community contributions.
GitHub Β· [npm](https://www.npmjs.com/package/resilient-queue
Backend Developer - Lagnify (Freelance) Β· Jun 2025 - Nov 2025
- Designed and built a 3-module REST API backend (auth, orders, admin) for lagnify.com using Node.js and MongoDB
- Owned the full deployment lifecycle on Render: DNS configuration, HTTPS provisioning, environment management, and production monitoring for reliable backend availability
B.Tech - Computer Engineering, JSPM RSCOE, Pune - 2023 - 2027 - CGPA: 8.95
- Reliability: expect failure, design for recovery, idempotency first
- Observability: log everything (structured), trace distributed systems, alert on anomaly
- Architecture: decouple with queues, retry transient failures, isolate fault zones
- Philosophy: if it's not tested, it's not working; build for recovery, not perfection; failure is expected
- LeetCode Knight | Contest Rating 1960+ | Top 3.05% Globally
- Weekly Contest 508 - Rank 361/37,970 (Top 0.95%)
- Recursion 3.0 - Runner-up
- Mayur Bhusare
"Reliable systems are built by engineers who expect failure."