Anshul Kumar

Full Stack & Distributed Systems Engineer

Email • LinkedIn • GitHub • Codeforces • LeetCode

---

About

I am an engineer focused on building scalable, high-performance backend systems and infrastructure. I enjoy tackling complex architectural problems, from designing custom database clones to developing robust remote code execution engines.

Currently pursuing my B.Tech in Computer Science Engineering (CGPA: 9.11/10.0) at Vellore Institute of Technology (VIT), Bhopal (Expected June 2027).

• Core Focus: Distributed systems, backend architecture, and algorithmic efficiency.
• Current Goal: Pushing high-throughput system boundaries and advancing in competitive programming.

Education

Vellore Institute of Technology (VIT), Bhopal — B.Tech in Computer Science and Engineering (Sep 2023 - Jun 2027)

• CGPA: 9.11/10.0
• Coursework: Object-Oriented Programming, Data Structures & Algorithms, Operating Systems, Computer Networks, DBMS, Distributed Systems, System Design.

Technical Expertise

• Languages: Java (JDK 21), C++ (STL), JavaScript (ES6+), Python, SQL, C
• Backend & Systems: Spring Boot 3, FastAPI, Node.js, Express, Docker, Microservices, Raw Sockets, Multithreading, TCP/IP, Concurrency (ConcurrentHashMap, ThreadPoolExecutor)
• Frontend & Web: React.js, Next.js, Node.js, Socket.io, Tailwind CSS, WebSockets, REST APIs
• Data Analytics & AI: Python (Pandas), Google Gemini API, ChromaDB (Vector DB), Recharts
• Tools: Git, GitHub, Linux, Postman, VS Code

Open Source Contributions

redis / jedis (Official Redis Java Client) | STATUS: OPEN PR #4645

• Pull Request: fix(cluster): prevent lock-inversion deadlock in MultiNodePipelineBase (#4557)
• Implemented a thread-safe fix in MultiNodePipelineBase to resolve a critical multi-threading lock-inversion deadlock (Issue #4557).
• Eliminated eager connection borrowing during command appending to break hold-and-wait conditions under high-concurrency multi-node environments.
• Authored a deterministic concurrency integration test using CyclicBarrier and strict connection mocks.

TheAlgorithms / Java (200k+ Star Open Source Algorithms Library) | STATUS: MERGED PR #7544

• Pull Request: Add Concurrent Merge Sort Implementation
• Engineered a high-performance multi-threaded Concurrent Merge Sort algorithm using Java's ForkJoinPool and recursive task splitting to maximize CPU utilization on multi-core architectures.

Featured & Technical Projects

1. Distributed Code Engine | Live Demo — Java, Spring Boot, Docker, React

• Architected a remote code evaluation API running Python, C++, and Java concurrently in isolated Docker containers with a 100% execution success rate under parallel load testing.
• Engineered asynchronous process management with Java ProcessBuilder, accelerating runtime pipelines to 130+ req/sec with sub-70 ms mean response latency.

2. MiniRedis (In-Memory Key-Value Store) | Live TCP Server — Java (JDK 21), Raw Sockets, Concurrency

• Engineered a concurrent in-memory key-value storage engine from scratch over raw TCP sockets, benchmarking at a peak throughput of 94,600+ ops/sec.
• Orchestrated thread-safe data consistency across 500 concurrent client connections using ConcurrentHashMap and fine-grained locking, stabilizing P99 latencies at 3.2 ms with a 0.00% error rate.

3. Sentinel AI | Live Demo — Python, FastAPI, ChromaDB, Google Gemini LLM, Docker

• Built an AI-powered observability platform that ingests production crash logs, performs semantic code retrieval via ChromaDB, and generates structured root-cause analyses and patch diffs with Gemini.
• Designed an asynchronous FastAPI telemetry pipeline with crash deduplication and vector search to reduce diagnosis latency for recurring production incidents.

4. ReconAI (AI-Powered Reconciliation Platform) | Live Demo — Python, FastAPI, React, Pandas, Gemini

• Designed an automated data reconciliation pipeline processing 5,000+ synthetic payment gateway transactions with a peak throughput of 420.16 req/sec and P99 latency of 38.63 ms.
• Automated root-cause analysis (RCA) on failed transaction batches by integrating the Gemini LLM, guaranteeing a 100% resolution success rate under concurrent load testing.

5. AI-Powered Customer Analytics Platform | Live Demo — Python, Streamlit, Random Forest, Gemini API

• Engineered an AI operational CRM and analytics dashboard to quantify and mitigate revenue at risk from payment gateway failures and technical debt.
• Implemented a Random Forest classifier to predict customer churn probability and utilized Gemini LLM to automatically draft personalized retention outreach campaigns.

6. SyncDraw | Live Demo — React.js, Node.js, Socket.io, WebSockets

• Built a collaborative whiteboard web application featuring zero-latency multi-user synchronization over a Pub/Sub WebSocket architecture.
• Engineered optimized Socket.io event broadcasting to transmit canvas coordinates instantly with minimal network payload overhead across connected clients.

7. L7 Load Balancer | Live Demo — C++, Multi-threading, TCP/IP, HTTP Server, Lock-free Atomics

• Engineered a custom L7 Reverse Proxy distributing traffic across backend servers using Round Robin and Least Connections routing algorithms.
• Utilized lock-free Atomic counters and multi-threaded worker pools to eliminate synchronization bottlenecks under concurrent connection spikes.

8. System Design Learning Platform | Live Demo — React.js, Node.js, Express, Tailwind CSS

• Developed a comprehensive interactive platform for engineering students to master scalable System Design architectures and distributed systems patterns.
• Structured real-world architectural blueprints, caching strategies, and database sharding paradigms into interactive visual study modules.

9. CodeLens AI | Live Demo — Python, FastAPI, React.js, Google Gemini Flash, Docker

• Built an automated code review agent that audits code snippets for time and space complexity bottlenecks, catches bugs, and generates refactored solutions.
• Designed an asynchronous inspection pipeline powered by Gemini Flash that delivers sub-second refactor recommendations and complexity breakdowns.

Achievements & Competitive Programming

• Codeforces: Attained Pupil rank (Peak Rating: 1289). Secured global rank 2185 in Educational CF Round 190.
• LeetCode: 1450 contest rating.
• Mastered 360+ DSA problems across both platforms using C++ (STL), resolving 118 Medium and Hard challenges.

Certifications

• AWS Certified Solutions Architect – Associate (Amazon Web Services)
• Networking Basics (Cisco)
• Cyber Security Analyst (IBM Career Education Program)
• Blockchain and its Applications (NPTEL / IIT Madras)