🎓 MS in Computer Science @ Northeastern University
⚙️ Systems & Infrastructure Engineer | Distributed Systems | Performance Engineering

I build high-performance, fault-tolerant systems with a focus on concurrency, low-latency messaging, streaming engines, and distributed storage.
My work emphasizes measurable performance, correctness under failure, and production-grade system design.

• Concurrent & Lock-Free Programming (C++ atomics, fine-grained locking)
• Distributed Systems (replication, partitioning, fault tolerance)
• Streaming & Messaging Systems (exactly-once semantics, backpressure)
• Performance Engineering (P99 latency, throughput benchmarking)
• Reliability (WAL, checkpoints, chaos testing, recovery)

Go · Kafka · BoltDB · Prometheus · Docker

• Achieved 163K events/sec with <20µs P99 latency using hash-partitioned parallel workers
• Implemented event-time windowing (tumbling, sliding, session) with millisecond precision
• Built checkpoint-replay system enabling exactly-once processing and <8s recovery
• Integrated token-bucket backpressure to sustain throughput under 10× traffic spikes

C++17 · Lock-Free Programming · mmap · Docker

• Built lock-free SPSC queue achieving 83ns P99 latency and 1.2M msgs/sec
• Reduced CPU utilization by 73% using zero-copy memory-mapped I/O
• Designed crash-recoverable WAL with CRC32 integrity for zero message loss
• Validated with extensive benchmarks, stress tests, and containerized deployment

Python · FastAPI · Consistent Hashing · Prometheus · Docker

• Architected 3-node distributed KV store delivering 250K+ ops/sec with <5ms P99 latency
• Implemented async replication with Merkle-tree anti-entropy, cutting recovery time by 85%
• Built heartbeat-based failover ensuring 99.9% availability during node failures
• Verified reliability using chaos testing and network partition simulations

Python · concurrent.futures · Docker · Linux

• Improved download throughput by 41.7% (1.71×) using HTTP Range request parallelism
• Designed deadlock-free, thread-safe file writer with zero corruption across 1,000+ concurrent writes
• Implemented exponential backoff retries improving success rate by 35% under packet loss
• Delivered modular, test-driven architecture with 48+ unit/integration/performance tests

Languages: Python, C++, C, Java, SQL, Bash
Databases: PostgreSQL, MySQL, MongoDB, SQLite, Redis
Systems & Cloud: Linux, Docker, AWS, GCP, FastAPI, Flask
Tools: Git, CMake, Jenkins, Prometheus, NumPy, PyTorch, TensorFlow

AI Intern — JustDial (AI Team)

• Built generative AI ad pipelines increasing engagement by 25% and CTR by 15%
• Optimized object detection pipelines achieving 92% accuracy and 30% lower latency
• Improved large-scale image processing workflows, reducing end-to-end processing time by 20%

• GitHub: https://github.com/yashpatil-byte
• LinkedIn: https://www.linkedin.com/in/yash-patil-57a53a1a6
• Email: yashpatilcareer@gmail.com

⭐ If you’re interested in systems engineering, distributed infrastructure, or performance-critical software, feel free to explore my projects or reach out.