Performance-Driven Developer • Systems Architect • Problem Solver

I specialize in bridging the gap between high-level application logic and low-level system efficiency. My work focuses on building tools that don't just work, but scale under pressure.

An enterprise-grade middleware designed to protect APIs from DoS and cascading failures.

• Performance: Optimized to handle 10k+ requests/sec with sub-millisecond overhead.
• Architectural Depth: Implemented Sliding Window Log & Token Bucket algorithms using Lua scripting in Redis to ensure atomicity.
• Impact: Solves the "race condition" problem inherent in basic distributed counters.
• [Repo Link] | [System Design Doc]

Custom storage engine engineered for analytical query speed.

• Optimization: Achieved up to 10x faster aggregation than row-based CSV parsing via columnar memory mapping.
• Data Compression: Integrated Dictionary Encoding and RLE, reducing disk footprint by ~40% for redundant datasets.
• [Technical Writeup] | [Benchmarks]

Deep dive into transformer architectures and real-time compliance monitoring.

• Generative AI: Built a GPT-style model from scratch to master Attention Mechanisms and Backpropagation.
• Edge AI: Developed a real-time Helmet Detection System using YOLO logic, optimized for low-latency inference on edge devices.
• [Project Demo] | [Model Metrics]

• DSA: 250+ LeetCode problems solved with a focus on Dynamic Programming and Graph Theory.
• CS Core: Deep understanding of Operating Systems (Process Sync, Deadlocks), Computer Networks (TCP/IP, OSI), and DBMS (Indexing, Sharding).
• Open Source: Active contributor to community-driven software, focusing on code quality and documentation.

I am currently seeking SDE / ML Engineer opportunities where I can contribute to high-stakes backend architecture or intelligent systems.

• 📍 Location: Punjab, India (Open to Remote/Relocation)
• 📧 Direct: amankumar005562@gmail.com
• 🔗 LinkedIn: /in/aman32
• 🌐 Portfolio: aman-portfolio-stable.vercel.app