Software Engineer focused on backend systems, Linux-based infrastructure, and low-latency software.

I enjoy building systems where performance, reliability, and correctness matter — from distributed backend services and real-time communication platforms to low-latency trading infrastructure and systems software.

Previously worked on backend systems, ERP/CRM platforms, online assessment infrastructure, and AI-powered applications using Go, PostgreSQL, Redis, RabbitMQ, Docker, and AWS.

• Building high-performance backend systems in Go and C++
• Exploring Linux networking, concurrency, and systems programming
• Contributing to open-source infrastructure projects
• Building low-latency systems and performance-focused infrastructure

C++ • Go • Python • JavaScript • Bash

Linux • POSIX Threads • Socket Programming • TCP/IP • epoll • Lock-Free Programming • Concurrent Systems

Go (Gin/Chi) • FastAPI • Node.js • Express.js • REST APIs • gRPC • WebSockets • RabbitMQ

PostgreSQL • Redis • MongoDB

Docker • AWS EC2 • Git • SSH • Neovim

FAISS • LangChain • LangGraph

C++20 • Linux Networking • epoll • Lock-Free Queues

• Built a complete trading ecosystem consisting of a Trading Firm Simulator, Exchange Gateway, Risk Engine, and Matching Engine.
• Implemented thread-per-shard ingress using Linux SO_REUSEPORT and epoll.
• Designed lock-free SPSC queues, memory pools, and cycle-accurate telemetry using RDTSCP.
• Sustained throughput exceeding 10 million messages per second.
• Demonstrated that approximately 77% of ingestion overhead originated from kernel networking rather than application business logic.

Repository: github.com/Kprateek283/High-Frequency-Trading-Engine

Go • Python • gRPC • PostgreSQL • Redis

• Built a distributed backend architecture for AI-driven NPC interactions.
• Implemented persistent memory, quest systems, and relationship tracking.
• Designed real-time communication using WebSockets and gRPC.
• Integrated retrieval-based and agent-based AI workflows.

Repository: github.com/Kprateek283/agentic-npc-project

Node.js • Docker • AWS

• Built a secure multi-language online judge supporting C, C++, Java, and Python.
• Implemented Docker-based sandbox isolation and resource constraints.
• Designed concurrent execution pipelines and automated evaluation workflows.

Repository: github.com/Kprateek283/OnlineJudge

Contributor to Linux io_uring userspace library.

• Upstreamed a multishot echo server example reviewed and merged by Jens Axboe.
• Additional contribution under review.

• Codeforces Expert (1706)
• Winner — CodeHunt 2025 (IIIT Manipur)
• Smart India Hackathon 2023 National Finalist
• NPTEL Top 1% — Software Testing
• NPTEL Top 1% — Industrial IoT
• NPTEL Top 5% — Data Analytics with Python
• NPTEL Top 5% — Human Computer Interaction

B.Tech, Computer Science & Engineering Indian Institute of Information Technology, Manipur

📧 Email: kprateek283@gmail.com

💼 LinkedIn: https://www.linkedin.com/in/prateek-kumar-400a14257

🏆 Codeforces: https://codeforces.com/profile/prats2853

💻 GitHub: https://github.com/Kprateek283

Interested in backend engineering, distributed systems, Linux internals, infrastructure, open source, and low-latency software.