I am a Software Engineer specializing in distributed systems, backend infrastructure, and database storage engines. Driven by first-principles engineering and mechanical sympathy, my focus spans low-level Linux systems programming and high-concurrency JVM internals up to fault-tolerant consensus invariants, LSM-tree storage engines, and event-driven architectures.
$ cat << 'EOF' > ethos.txt
"Design with mechanical sympathy for the hardware,
zero-cost abstraction for the runtime,
and rigorous fault tolerance for the network."
EOF- 🏛️ Architecture & Scale: Proven track record architecting event-driven platforms processing 500,000+ events/sec with zero-copy I/O.
- 💾 Storage Internals: Hands-on implementer of Write-Ahead Logging (WAL), MemTable SkipLists, binary SSTables, and background compaction algorithms.
- ☕ High-Concurrency Runtimes: Deep expertise in JVM internals, Java 21 Project Loom virtual threads, JMM memory barriers, and lock-free CAS primitives.
- 📚 Engineering Hub: Author of the 2026 Master Systems Architecture & Study Hub, synthesizing 800+ technical guides, 509 Markdown docs, and 283 Java masterclasses.
| Project | Architecture & Highlights | Key Metrics |
|---|---|---|
|
🦀 MiniDB Storage Engine Embedded LSM-Tree Key-Value Engine Rust LSM-Tree WAL SSTable
|
High-performance embedded storage engine implemented in Rust. Features append-only Write-Ahead Logging (WAL) for durability, lock-free MemTable SkipList for high write ingestion, binary SSTable layout with Bloom Filters, and background leveled compaction. |
• P99 Write: < 0.8ms • Bloom Filter: 1% FPR • Zero Data Loss on crash |
|
⚡ Twitter Snowflake ID Generator Distributed 64-bit ID Generation Java 21 Lock-Free CAS Atomic
|
High-concurrency, coordinate-free unique 64-bit ID generator based on Twitter Snowflake. Eliminates cross-datacenter network consensus while maintaining strict time-sortability and sub-millisecond sequence rollover with NTP clock-skew protection. |
• 4,096 IDs/ms per worker • Zero consensus rounds • Lock-Free atomic bit shifts |
|
🌊 Streaming Pipeline Engine Zero-Copy Event Ingestion Architecture Kafka Linux Kernel sendfile()
|
High-throughput event streaming pipeline utilizing Apache Kafka, partitioned consumer group rebalances, and Linux kernel PageCache socket transfers via sendfile(). Implements Transactional Outbox and idempotent consumer semantics.
|
• 500k+ events/sec • Zero-Copy socket I/O • Exactly-Once guarantees |
|
🧠 2026 Master Study & Prep Hub Systems Engineering & Architecture Platform Distributed Systems DDIA OSTEP
|
A production-grade, 13-week curriculum covering 800+ technical modules. Synthesizes canonical textbooks (DDIA, OSTEP, Effective Java, Database Internals), university courses (MIT 6.5840, Stanford CS244B), and 28 real-world Alex Xu high-level design systems. |
• 800+ technical guides • 509 deep-dive docs • 283 Java masterclasses |
|
🏛️ Production LLD Engine Suite Production Low-Level Design Engine Design Patterns SOLID Concurrency
|
Production implementations of core systems: Thread-safe multi-threaded LRU Cache with MPSC write buffers, concurrent Elevator Dispatching Engine, Multi-Floor Parking Lot, and distributed Token Bucket Rate Limiters with ArchUnit boundary verification. |
• 5 full case studies • Hexagonal architecture • Thread-Safe invariants |
|
|
|
|
My engineering framework is grounded in foundational computer science literature and canonical university curricula:
- 🎓 MIT 6.5840 (6.824): Distributed Systems (MapReduce, Raft Consensus, Primary-Backup, 2PC)
- 🎓 Stanford CS244B: Advanced Distributed Systems (P2P DHTs, Byzantine Fault Tolerance, Quorum Replication)
- 📘 DDIA (Martin Kleppmann): Designing Data-Intensive Applications (SSTables, Replication Lag, Linearizability)
- 📗 OSTEP (Remzi & Andrea Arpaci-Dusseau): Operating Systems: Three Easy Pieces (Virtualization, Concurrency, File Systems)
- 📙 Database Internals (Alex Petrov): B-Trees, LSM-Trees, Page Cache, and Distributed Storage Engines
|
|
|
|
|
|
⚡ Engineered with mechanical sympathy by Kartik Chauhan